diff --git a/.git-blame-ignore-revs b/.git-blame-ignore-revs index 1ee1dee011164..f83e2fc658664 100644 --- a/.git-blame-ignore-revs +++ b/.git-blame-ignore-revs @@ -14,6 +14,12 @@ # # $ git log --pretty=format:"%H # %cd%n# %s" $PGINDENTGITHASH -1 --date=iso +1d1612aec7688139e1a5506df1366b4b6a69605d # 2025-07-29 09:10:41 -0400 +# Run pgindent. + +73873805fb3627cb23937c750fa83ffd8f16fc6c # 2025-07-25 16:36:44 -0400 +# Run pgindent on the changes of the previous patch. + 9e345415bcd3c4358350b89edfd710469b8bfaf9 # 2025-07-01 15:23:07 +0200 # Fix indentation in pg_numa code diff --git a/config/llvm.m4 b/config/llvm.m4 index fa4bedd9370fc..9d6fe8199e364 100644 --- a/config/llvm.m4 +++ b/config/llvm.m4 @@ -4,7 +4,7 @@ # ----------------- # # Look for the LLVM installation, check that it's new enough, set the -# corresponding LLVM_{CFLAGS,CXXFLAGS,BINPATH} and LDFLAGS +# corresponding LLVM_{CFLAGS,CXXFLAGS,BINPATH,LIBS} # variables. Also verify that CLANG is available, to transform C # into bitcode. # @@ -55,7 +55,7 @@ AC_DEFUN([PGAC_LLVM_SUPPORT], for pgac_option in `$LLVM_CONFIG --ldflags`; do case $pgac_option in - -L*) LDFLAGS="$LDFLAGS $pgac_option";; + -L*) LLVM_LIBS="$LLVM_LIBS $pgac_option";; esac done diff --git a/config/programs.m4 b/config/programs.m4 index c73d9307ea8a9..e57fe4907b844 100644 --- a/config/programs.m4 +++ b/config/programs.m4 @@ -290,8 +290,8 @@ AC_DEFUN([PGAC_CHECK_LIBCURL], pgac_save_LDFLAGS=$LDFLAGS pgac_save_LIBS=$LIBS - CPPFLAGS="$LIBCURL_CPPFLAGS $CPPFLAGS" - LDFLAGS="$LIBCURL_LDFLAGS $LDFLAGS" + CPPFLAGS="$CPPFLAGS $LIBCURL_CPPFLAGS" + LDFLAGS="$LDFLAGS $LIBCURL_LDFLAGS" AC_CHECK_HEADER(curl/curl.h, [], [AC_MSG_ERROR([header file is required for --with-libcurl])]) diff --git a/configure b/configure index 6d7c22e153fea..507a2437c3308 100755 --- a/configure +++ b/configure @@ -5194,7 +5194,7 @@ fi for pgac_option in `$LLVM_CONFIG --ldflags`; do case $pgac_option in - -L*) LDFLAGS="$LDFLAGS $pgac_option";; + -L*) LLVM_LIBS="$LLVM_LIBS $pgac_option";; esac done @@ -9436,12 +9436,12 @@ fi # Note the user could also set XML2_CFLAGS/XML2_LIBS directly for pgac_option in $XML2_CFLAGS; do case $pgac_option in - -I*|-D*) CPPFLAGS="$CPPFLAGS $pgac_option";; + -I*|-D*) INCLUDES="$INCLUDES $pgac_option";; esac done for pgac_option in $XML2_LIBS; do case $pgac_option in - -L*) LDFLAGS="$LDFLAGS $pgac_option";; + -L*) LIBDIRS="$LIBDIRS $pgac_option";; esac done fi @@ -9666,12 +9666,12 @@ fi # note that -llz4 will be added by AC_CHECK_LIB below. for pgac_option in $LZ4_CFLAGS; do case $pgac_option in - -I*|-D*) CPPFLAGS="$CPPFLAGS $pgac_option";; + -I*|-D*) INCLUDES="$INCLUDES $pgac_option";; esac done for pgac_option in $LZ4_LIBS; do case $pgac_option in - -L*) LDFLAGS="$LDFLAGS $pgac_option";; + -L*) LIBDIRS="$LIBDIRS $pgac_option";; esac done fi @@ -9807,12 +9807,12 @@ fi # note that -lzstd will be added by AC_CHECK_LIB below. for pgac_option in $ZSTD_CFLAGS; do case $pgac_option in - -I*|-D*) CPPFLAGS="$CPPFLAGS $pgac_option";; + -I*|-D*) INCLUDES="$INCLUDES $pgac_option";; esac done for pgac_option in $ZSTD_LIBS; do case $pgac_option in - -L*) LDFLAGS="$LDFLAGS $pgac_option";; + -L*) LIBDIRS="$LIBDIRS $pgac_option";; esac done fi @@ -12723,8 +12723,8 @@ if test "$with_libcurl" = yes ; then pgac_save_LDFLAGS=$LDFLAGS pgac_save_LIBS=$LIBS - CPPFLAGS="$LIBCURL_CPPFLAGS $CPPFLAGS" - LDFLAGS="$LIBCURL_LDFLAGS $LDFLAGS" + CPPFLAGS="$CPPFLAGS $LIBCURL_CPPFLAGS" + LDFLAGS="$LDFLAGS $LIBCURL_LDFLAGS" ac_fn_c_check_header_mongrel "$LINENO" "curl/curl.h" "ac_cv_header_curl_curl_h" "$ac_includes_default" if test "x$ac_cv_header_curl_curl_h" = xyes; then : @@ -16658,7 +16658,7 @@ fi if test "$with_icu" = yes; then ac_save_CPPFLAGS=$CPPFLAGS - CPPFLAGS="$ICU_CFLAGS $CPPFLAGS" + CPPFLAGS="$CPPFLAGS $ICU_CFLAGS" # Verify we have ICU's header files ac_fn_c_check_header_mongrel "$LINENO" "unicode/ucol.h" "ac_cv_header_unicode_ucol_h" "$ac_includes_default" @@ -17565,7 +17565,7 @@ $as_echo "#define HAVE_GCC__ATOMIC_INT64_CAS 1" >>confdefs.h fi -# Check for x86 cpuid instruction +# Check for __get_cpuid() and __cpuid() { $as_echo "$as_me:${as_lineno-$LINENO}: checking for __get_cpuid" >&5 $as_echo_n "checking for __get_cpuid... " >&6; } if ${pgac_cv__get_cpuid+:} false; then : @@ -17598,77 +17598,79 @@ if test x"$pgac_cv__get_cpuid" = x"yes"; then $as_echo "#define HAVE__GET_CPUID 1" >>confdefs.h -fi - -{ $as_echo "$as_me:${as_lineno-$LINENO}: checking for __get_cpuid_count" >&5 -$as_echo_n "checking for __get_cpuid_count... " >&6; } -if ${pgac_cv__get_cpuid_count+:} false; then : +else + # __cpuid() + { $as_echo "$as_me:${as_lineno-$LINENO}: checking for __cpuid" >&5 +$as_echo_n "checking for __cpuid... " >&6; } +if ${pgac_cv__cpuid+:} false; then : $as_echo_n "(cached) " >&6 else cat confdefs.h - <<_ACEOF >conftest.$ac_ext /* end confdefs.h. */ -#include +#include int main () { unsigned int exx[4] = {0, 0, 0, 0}; - __get_cpuid_count(7, 0, &exx[0], &exx[1], &exx[2], &exx[3]); + __cpuid(exx, 1); ; return 0; } _ACEOF if ac_fn_c_try_link "$LINENO"; then : - pgac_cv__get_cpuid_count="yes" + pgac_cv__cpuid="yes" else - pgac_cv__get_cpuid_count="no" + pgac_cv__cpuid="no" fi rm -f core conftest.err conftest.$ac_objext \ conftest$ac_exeext conftest.$ac_ext fi -{ $as_echo "$as_me:${as_lineno-$LINENO}: result: $pgac_cv__get_cpuid_count" >&5 -$as_echo "$pgac_cv__get_cpuid_count" >&6; } -if test x"$pgac_cv__get_cpuid_count" = x"yes"; then +{ $as_echo "$as_me:${as_lineno-$LINENO}: result: $pgac_cv__cpuid" >&5 +$as_echo "$pgac_cv__cpuid" >&6; } + if test x"$pgac_cv__cpuid" = x"yes"; then -$as_echo "#define HAVE__GET_CPUID_COUNT 1" >>confdefs.h +$as_echo "#define HAVE__CPUID 1" >>confdefs.h + fi fi -{ $as_echo "$as_me:${as_lineno-$LINENO}: checking for __cpuid" >&5 -$as_echo_n "checking for __cpuid... " >&6; } -if ${pgac_cv__cpuid+:} false; then : +# Check for __get_cpuid_count() and __cpuidex() in a similar fashion. +{ $as_echo "$as_me:${as_lineno-$LINENO}: checking for __get_cpuid_count" >&5 +$as_echo_n "checking for __get_cpuid_count... " >&6; } +if ${pgac_cv__get_cpuid_count+:} false; then : $as_echo_n "(cached) " >&6 else cat confdefs.h - <<_ACEOF >conftest.$ac_ext /* end confdefs.h. */ -#include +#include int main () { unsigned int exx[4] = {0, 0, 0, 0}; - __get_cpuid(exx[0], 1); + __get_cpuid_count(7, 0, &exx[0], &exx[1], &exx[2], &exx[3]); ; return 0; } _ACEOF if ac_fn_c_try_link "$LINENO"; then : - pgac_cv__cpuid="yes" + pgac_cv__get_cpuid_count="yes" else - pgac_cv__cpuid="no" + pgac_cv__get_cpuid_count="no" fi rm -f core conftest.err conftest.$ac_objext \ conftest$ac_exeext conftest.$ac_ext fi -{ $as_echo "$as_me:${as_lineno-$LINENO}: result: $pgac_cv__cpuid" >&5 -$as_echo "$pgac_cv__cpuid" >&6; } -if test x"$pgac_cv__cpuid" = x"yes"; then - -$as_echo "#define HAVE__CPUID 1" >>confdefs.h +{ $as_echo "$as_me:${as_lineno-$LINENO}: result: $pgac_cv__get_cpuid_count" >&5 +$as_echo "$pgac_cv__get_cpuid_count" >&6; } +if test x"$pgac_cv__get_cpuid_count" = x"yes"; then -fi +$as_echo "#define HAVE__GET_CPUID_COUNT 1" >>confdefs.h -{ $as_echo "$as_me:${as_lineno-$LINENO}: checking for __cpuidex" >&5 +else + # __cpuidex() + { $as_echo "$as_me:${as_lineno-$LINENO}: checking for __cpuidex" >&5 $as_echo_n "checking for __cpuidex... " >&6; } if ${pgac_cv__cpuidex+:} false; then : $as_echo_n "(cached) " >&6 @@ -17680,7 +17682,7 @@ int main () { unsigned int exx[4] = {0, 0, 0, 0}; - __get_cpuidex(exx[0], 7, 0); + __cpuidex(exx, 7, 0); ; return 0; @@ -17696,10 +17698,11 @@ rm -f core conftest.err conftest.$ac_objext \ fi { $as_echo "$as_me:${as_lineno-$LINENO}: result: $pgac_cv__cpuidex" >&5 $as_echo "$pgac_cv__cpuidex" >&6; } -if test x"$pgac_cv__cpuidex" = x"yes"; then + if test x"$pgac_cv__cpuidex" = x"yes"; then $as_echo "#define HAVE__CPUIDEX 1" >>confdefs.h + fi fi # Check for XSAVE intrinsics @@ -18876,7 +18879,7 @@ Use --without-tcl to disable building PL/Tcl." "$LINENO" 5 fi # now that we have TCL_INCLUDE_SPEC, we can check for ac_save_CPPFLAGS=$CPPFLAGS - CPPFLAGS="$TCL_INCLUDE_SPEC $CPPFLAGS" + CPPFLAGS="$CPPFLAGS $TCL_INCLUDE_SPEC" ac_fn_c_check_header_mongrel "$LINENO" "tcl.h" "ac_cv_header_tcl_h" "$ac_includes_default" if test "x$ac_cv_header_tcl_h" = xyes; then : @@ -18945,7 +18948,7 @@ fi # check for if test "$with_python" = yes; then ac_save_CPPFLAGS=$CPPFLAGS - CPPFLAGS="$python_includespec $CPPFLAGS" + CPPFLAGS="$CPPFLAGS $python_includespec" ac_fn_c_check_header_mongrel "$LINENO" "Python.h" "ac_cv_header_Python_h" "$ac_includes_default" if test "x$ac_cv_header_Python_h" = xyes; then : diff --git a/configure.ac b/configure.ac index c2877e369350e..5f4548adc5cd1 100644 --- a/configure.ac +++ b/configure.ac @@ -1103,12 +1103,12 @@ if test "$with_libxml" = yes ; then # Note the user could also set XML2_CFLAGS/XML2_LIBS directly for pgac_option in $XML2_CFLAGS; do case $pgac_option in - -I*|-D*) CPPFLAGS="$CPPFLAGS $pgac_option";; + -I*|-D*) INCLUDES="$INCLUDES $pgac_option";; esac done for pgac_option in $XML2_LIBS; do case $pgac_option in - -L*) LDFLAGS="$LDFLAGS $pgac_option";; + -L*) LIBDIRS="$LIBDIRS $pgac_option";; esac done fi @@ -1152,12 +1152,12 @@ if test "$with_lz4" = yes; then # note that -llz4 will be added by AC_CHECK_LIB below. for pgac_option in $LZ4_CFLAGS; do case $pgac_option in - -I*|-D*) CPPFLAGS="$CPPFLAGS $pgac_option";; + -I*|-D*) INCLUDES="$INCLUDES $pgac_option";; esac done for pgac_option in $LZ4_LIBS; do case $pgac_option in - -L*) LDFLAGS="$LDFLAGS $pgac_option";; + -L*) LIBDIRS="$LIBDIRS $pgac_option";; esac done fi @@ -1177,12 +1177,12 @@ if test "$with_zstd" = yes; then # note that -lzstd will be added by AC_CHECK_LIB below. for pgac_option in $ZSTD_CFLAGS; do case $pgac_option in - -I*|-D*) CPPFLAGS="$CPPFLAGS $pgac_option";; + -I*|-D*) INCLUDES="$INCLUDES $pgac_option";; esac done for pgac_option in $ZSTD_LIBS; do case $pgac_option in - -L*) LDFLAGS="$LDFLAGS $pgac_option";; + -L*) LIBDIRS="$LIBDIRS $pgac_option";; esac done fi @@ -1944,7 +1944,7 @@ fi if test "$with_icu" = yes; then ac_save_CPPFLAGS=$CPPFLAGS - CPPFLAGS="$ICU_CFLAGS $CPPFLAGS" + CPPFLAGS="$CPPFLAGS $ICU_CFLAGS" # Verify we have ICU's header files AC_CHECK_HEADER(unicode/ucol.h, [], @@ -2044,7 +2044,7 @@ PGAC_HAVE_GCC__ATOMIC_INT32_CAS PGAC_HAVE_GCC__ATOMIC_INT64_CAS -# Check for x86 cpuid instruction +# Check for __get_cpuid() and __cpuid() AC_CACHE_CHECK([for __get_cpuid], [pgac_cv__get_cpuid], [AC_LINK_IFELSE([AC_LANG_PROGRAM([#include ], [[unsigned int exx[4] = {0, 0, 0, 0}; @@ -2054,8 +2054,21 @@ AC_CACHE_CHECK([for __get_cpuid], [pgac_cv__get_cpuid], [pgac_cv__get_cpuid="no"])]) if test x"$pgac_cv__get_cpuid" = x"yes"; then AC_DEFINE(HAVE__GET_CPUID, 1, [Define to 1 if you have __get_cpuid.]) +else + # __cpuid() + AC_CACHE_CHECK([for __cpuid], [pgac_cv__cpuid], + [AC_LINK_IFELSE([AC_LANG_PROGRAM([#include ], + [[unsigned int exx[4] = {0, 0, 0, 0}; + __cpuid(exx, 1); + ]])], + [pgac_cv__cpuid="yes"], + [pgac_cv__cpuid="no"])]) + if test x"$pgac_cv__cpuid" = x"yes"; then + AC_DEFINE(HAVE__CPUID, 1, [Define to 1 if you have __cpuid.]) + fi fi +# Check for __get_cpuid_count() and __cpuidex() in a similar fashion. AC_CACHE_CHECK([for __get_cpuid_count], [pgac_cv__get_cpuid_count], [AC_LINK_IFELSE([AC_LANG_PROGRAM([#include ], [[unsigned int exx[4] = {0, 0, 0, 0}; @@ -2065,28 +2078,18 @@ AC_CACHE_CHECK([for __get_cpuid_count], [pgac_cv__get_cpuid_count], [pgac_cv__get_cpuid_count="no"])]) if test x"$pgac_cv__get_cpuid_count" = x"yes"; then AC_DEFINE(HAVE__GET_CPUID_COUNT, 1, [Define to 1 if you have __get_cpuid_count.]) -fi - -AC_CACHE_CHECK([for __cpuid], [pgac_cv__cpuid], -[AC_LINK_IFELSE([AC_LANG_PROGRAM([#include ], - [[unsigned int exx[4] = {0, 0, 0, 0}; - __get_cpuid(exx[0], 1); - ]])], - [pgac_cv__cpuid="yes"], - [pgac_cv__cpuid="no"])]) -if test x"$pgac_cv__cpuid" = x"yes"; then - AC_DEFINE(HAVE__CPUID, 1, [Define to 1 if you have __cpuid.]) -fi - -AC_CACHE_CHECK([for __cpuidex], [pgac_cv__cpuidex], -[AC_LINK_IFELSE([AC_LANG_PROGRAM([#include ], - [[unsigned int exx[4] = {0, 0, 0, 0}; - __get_cpuidex(exx[0], 7, 0); - ]])], - [pgac_cv__cpuidex="yes"], - [pgac_cv__cpuidex="no"])]) -if test x"$pgac_cv__cpuidex" = x"yes"; then - AC_DEFINE(HAVE__CPUIDEX, 1, [Define to 1 if you have __cpuidex.]) +else + # __cpuidex() + AC_CACHE_CHECK([for __cpuidex], [pgac_cv__cpuidex], + [AC_LINK_IFELSE([AC_LANG_PROGRAM([#include ], + [[unsigned int exx[4] = {0, 0, 0, 0}; + __cpuidex(exx, 7, 0); + ]])], + [pgac_cv__cpuidex="yes"], + [pgac_cv__cpuidex="no"])]) + if test x"$pgac_cv__cpuidex" = x"yes"; then + AC_DEFINE(HAVE__CPUIDEX, 1, [Define to 1 if you have __cpuidex.]) + fi fi # Check for XSAVE intrinsics @@ -2344,7 +2347,7 @@ Use --without-tcl to disable building PL/Tcl.]) fi # now that we have TCL_INCLUDE_SPEC, we can check for ac_save_CPPFLAGS=$CPPFLAGS - CPPFLAGS="$TCL_INCLUDE_SPEC $CPPFLAGS" + CPPFLAGS="$CPPFLAGS $TCL_INCLUDE_SPEC" AC_CHECK_HEADER(tcl.h, [], [AC_MSG_ERROR([header file is required for Tcl])]) CPPFLAGS=$ac_save_CPPFLAGS fi @@ -2381,7 +2384,7 @@ fi # check for if test "$with_python" = yes; then ac_save_CPPFLAGS=$CPPFLAGS - CPPFLAGS="$python_includespec $CPPFLAGS" + CPPFLAGS="$CPPFLAGS $python_includespec" AC_CHECK_HEADER(Python.h, [], [AC_MSG_ERROR([header file is required for Python])]) CPPFLAGS=$ac_save_CPPFLAGS fi diff --git a/contrib/auto_explain/Makefile b/contrib/auto_explain/Makefile index efd127d3cae64..94ab28e7c06b9 100644 --- a/contrib/auto_explain/Makefile +++ b/contrib/auto_explain/Makefile @@ -6,6 +6,8 @@ OBJS = \ auto_explain.o PGFILEDESC = "auto_explain - logging facility for execution plans" +REGRESS = alter_reset + TAP_TESTS = 1 ifdef USE_PGXS diff --git a/contrib/auto_explain/expected/alter_reset.out b/contrib/auto_explain/expected/alter_reset.out new file mode 100644 index 0000000000000..ec355189806ae --- /dev/null +++ b/contrib/auto_explain/expected/alter_reset.out @@ -0,0 +1,19 @@ +-- +-- This tests resetting unknown custom GUCs with reserved prefixes. There's +-- nothing specific to auto_explain; this is just a convenient place to put +-- this test. +-- +SELECT current_database() AS datname \gset +CREATE ROLE regress_ae_role; +ALTER DATABASE :"datname" SET auto_explain.bogus = 1; +ALTER ROLE regress_ae_role SET auto_explain.bogus = 1; +ALTER ROLE regress_ae_role IN DATABASE :"datname" SET auto_explain.bogus = 1; +ALTER SYSTEM SET auto_explain.bogus = 1; +LOAD 'auto_explain'; +WARNING: invalid configuration parameter name "auto_explain.bogus", removing it +DETAIL: "auto_explain" is now a reserved prefix. +ALTER DATABASE :"datname" RESET auto_explain.bogus; +ALTER ROLE regress_ae_role RESET auto_explain.bogus; +ALTER ROLE regress_ae_role IN DATABASE :"datname" RESET auto_explain.bogus; +ALTER SYSTEM RESET auto_explain.bogus; +DROP ROLE regress_ae_role; diff --git a/contrib/auto_explain/meson.build b/contrib/auto_explain/meson.build index 92dc9df6f7cac..a9b45cc235f12 100644 --- a/contrib/auto_explain/meson.build +++ b/contrib/auto_explain/meson.build @@ -20,6 +20,11 @@ tests += { 'name': 'auto_explain', 'sd': meson.current_source_dir(), 'bd': meson.current_build_dir(), + 'regress': { + 'sql': [ + 'alter_reset', + ], + }, 'tap': { 'tests': [ 't/001_auto_explain.pl', diff --git a/contrib/auto_explain/sql/alter_reset.sql b/contrib/auto_explain/sql/alter_reset.sql new file mode 100644 index 0000000000000..bf621454ec24a --- /dev/null +++ b/contrib/auto_explain/sql/alter_reset.sql @@ -0,0 +1,22 @@ +-- +-- This tests resetting unknown custom GUCs with reserved prefixes. There's +-- nothing specific to auto_explain; this is just a convenient place to put +-- this test. +-- + +SELECT current_database() AS datname \gset +CREATE ROLE regress_ae_role; + +ALTER DATABASE :"datname" SET auto_explain.bogus = 1; +ALTER ROLE regress_ae_role SET auto_explain.bogus = 1; +ALTER ROLE regress_ae_role IN DATABASE :"datname" SET auto_explain.bogus = 1; +ALTER SYSTEM SET auto_explain.bogus = 1; + +LOAD 'auto_explain'; + +ALTER DATABASE :"datname" RESET auto_explain.bogus; +ALTER ROLE regress_ae_role RESET auto_explain.bogus; +ALTER ROLE regress_ae_role IN DATABASE :"datname" RESET auto_explain.bogus; +ALTER SYSTEM RESET auto_explain.bogus; + +DROP ROLE regress_ae_role; diff --git a/contrib/dblink/dblink.c b/contrib/dblink/dblink.c index de5bed282f3f0..f98805fb5f735 100644 --- a/contrib/dblink/dblink.c +++ b/contrib/dblink/dblink.c @@ -101,8 +101,8 @@ static void materializeQueryResult(FunctionCallInfo fcinfo, const char *conname, const char *sql, bool fail); -static PGresult *storeQueryResult(volatile storeInfo *sinfo, PGconn *conn, const char *sql); -static void storeRow(volatile storeInfo *sinfo, PGresult *res, bool first); +static PGresult *storeQueryResult(storeInfo *sinfo, PGconn *conn, const char *sql); +static void storeRow(storeInfo *sinfo, PGresult *res, bool first); static remoteConn *getConnectionByName(const char *name); static HTAB *createConnHash(void); static remoteConn *createNewConnection(const char *name); @@ -169,14 +169,6 @@ typedef struct remoteConnHashEnt /* initial number of connection hashes */ #define NUMCONN 16 -static char * -xpstrdup(const char *in) -{ - if (in == NULL) - return NULL; - return pstrdup(in); -} - pg_noreturn static void dblink_res_internalerror(PGconn *conn, PGresult *res, const char *p2) { @@ -870,131 +862,123 @@ static void materializeResult(FunctionCallInfo fcinfo, PGconn *conn, PGresult *res) { ReturnSetInfo *rsinfo = (ReturnSetInfo *) fcinfo->resultinfo; + TupleDesc tupdesc; + bool is_sql_cmd; + int ntuples; + int nfields; /* prepTuplestoreResult must have been called previously */ Assert(rsinfo->returnMode == SFRM_Materialize); - PG_TRY(); + if (PQresultStatus(res) == PGRES_COMMAND_OK) { - TupleDesc tupdesc; - bool is_sql_cmd; - int ntuples; - int nfields; + is_sql_cmd = true; - if (PQresultStatus(res) == PGRES_COMMAND_OK) - { - is_sql_cmd = true; + /* + * need a tuple descriptor representing one TEXT column to return the + * command status string as our result tuple + */ + tupdesc = CreateTemplateTupleDesc(1); + TupleDescInitEntry(tupdesc, (AttrNumber) 1, "status", + TEXTOID, -1, 0); + ntuples = 1; + nfields = 1; + } + else + { + Assert(PQresultStatus(res) == PGRES_TUPLES_OK); - /* - * need a tuple descriptor representing one TEXT column to return - * the command status string as our result tuple - */ - tupdesc = CreateTemplateTupleDesc(1); - TupleDescInitEntry(tupdesc, (AttrNumber) 1, "status", - TEXTOID, -1, 0); - ntuples = 1; - nfields = 1; - } - else - { - Assert(PQresultStatus(res) == PGRES_TUPLES_OK); + is_sql_cmd = false; - is_sql_cmd = false; + /* get a tuple descriptor for our result type */ + switch (get_call_result_type(fcinfo, NULL, &tupdesc)) + { + case TYPEFUNC_COMPOSITE: + /* success */ + break; + case TYPEFUNC_RECORD: + /* failed to determine actual type of RECORD */ + ereport(ERROR, + (errcode(ERRCODE_FEATURE_NOT_SUPPORTED), + errmsg("function returning record called in context " + "that cannot accept type record"))); + break; + default: + /* result type isn't composite */ + elog(ERROR, "return type must be a row type"); + break; + } - /* get a tuple descriptor for our result type */ - switch (get_call_result_type(fcinfo, NULL, &tupdesc)) - { - case TYPEFUNC_COMPOSITE: - /* success */ - break; - case TYPEFUNC_RECORD: - /* failed to determine actual type of RECORD */ - ereport(ERROR, - (errcode(ERRCODE_FEATURE_NOT_SUPPORTED), - errmsg("function returning record called in context " - "that cannot accept type record"))); - break; - default: - /* result type isn't composite */ - elog(ERROR, "return type must be a row type"); - break; - } + /* make sure we have a persistent copy of the tupdesc */ + tupdesc = CreateTupleDescCopy(tupdesc); + ntuples = PQntuples(res); + nfields = PQnfields(res); + } - /* make sure we have a persistent copy of the tupdesc */ - tupdesc = CreateTupleDescCopy(tupdesc); - ntuples = PQntuples(res); - nfields = PQnfields(res); - } + /* + * check result and tuple descriptor have the same number of columns + */ + if (nfields != tupdesc->natts) + ereport(ERROR, + (errcode(ERRCODE_DATATYPE_MISMATCH), + errmsg("remote query result rowtype does not match " + "the specified FROM clause rowtype"))); - /* - * check result and tuple descriptor have the same number of columns - */ - if (nfields != tupdesc->natts) - ereport(ERROR, - (errcode(ERRCODE_DATATYPE_MISMATCH), - errmsg("remote query result rowtype does not match " - "the specified FROM clause rowtype"))); + if (ntuples > 0) + { + AttInMetadata *attinmeta; + int nestlevel = -1; + Tuplestorestate *tupstore; + MemoryContext oldcontext; + int row; + char **values; - if (ntuples > 0) - { - AttInMetadata *attinmeta; - int nestlevel = -1; - Tuplestorestate *tupstore; - MemoryContext oldcontext; - int row; - char **values; + attinmeta = TupleDescGetAttInMetadata(tupdesc); - attinmeta = TupleDescGetAttInMetadata(tupdesc); + /* Set GUCs to ensure we read GUC-sensitive data types correctly */ + if (!is_sql_cmd) + nestlevel = applyRemoteGucs(conn); - /* Set GUCs to ensure we read GUC-sensitive data types correctly */ - if (!is_sql_cmd) - nestlevel = applyRemoteGucs(conn); + oldcontext = MemoryContextSwitchTo(rsinfo->econtext->ecxt_per_query_memory); + tupstore = tuplestore_begin_heap(true, false, work_mem); + rsinfo->setResult = tupstore; + rsinfo->setDesc = tupdesc; + MemoryContextSwitchTo(oldcontext); - oldcontext = MemoryContextSwitchTo(rsinfo->econtext->ecxt_per_query_memory); - tupstore = tuplestore_begin_heap(true, false, work_mem); - rsinfo->setResult = tupstore; - rsinfo->setDesc = tupdesc; - MemoryContextSwitchTo(oldcontext); + values = palloc_array(char *, nfields); - values = palloc_array(char *, nfields); + /* put all tuples into the tuplestore */ + for (row = 0; row < ntuples; row++) + { + HeapTuple tuple; - /* put all tuples into the tuplestore */ - for (row = 0; row < ntuples; row++) + if (!is_sql_cmd) { - HeapTuple tuple; + int i; - if (!is_sql_cmd) - { - int i; - - for (i = 0; i < nfields; i++) - { - if (PQgetisnull(res, row, i)) - values[i] = NULL; - else - values[i] = PQgetvalue(res, row, i); - } - } - else + for (i = 0; i < nfields; i++) { - values[0] = PQcmdStatus(res); + if (PQgetisnull(res, row, i)) + values[i] = NULL; + else + values[i] = PQgetvalue(res, row, i); } - - /* build the tuple and put it into the tuplestore. */ - tuple = BuildTupleFromCStrings(attinmeta, values); - tuplestore_puttuple(tupstore, tuple); + } + else + { + values[0] = PQcmdStatus(res); } - /* clean up GUC settings, if we changed any */ - restoreLocalGucs(nestlevel); + /* build the tuple and put it into the tuplestore. */ + tuple = BuildTupleFromCStrings(attinmeta, values); + tuplestore_puttuple(tupstore, tuple); } + + /* clean up GUC settings, if we changed any */ + restoreLocalGucs(nestlevel); } - PG_FINALLY(); - { - /* be sure to release the libpq result */ - PQclear(res); - } - PG_END_TRY(); + + PQclear(res); } /* @@ -1013,16 +997,17 @@ materializeQueryResult(FunctionCallInfo fcinfo, bool fail) { ReturnSetInfo *rsinfo = (ReturnSetInfo *) fcinfo->resultinfo; - PGresult *volatile res = NULL; - volatile storeInfo sinfo = {0}; /* prepTuplestoreResult must have been called previously */ Assert(rsinfo->returnMode == SFRM_Materialize); - sinfo.fcinfo = fcinfo; - + /* Use a PG_TRY block to ensure we pump libpq dry of results */ PG_TRY(); { + storeInfo sinfo = {0}; + PGresult *res; + + sinfo.fcinfo = fcinfo; /* Create short-lived memory context for data conversions */ sinfo.tmpcontext = AllocSetContextCreate(CurrentMemoryContext, "dblink temporary context", @@ -1035,14 +1020,7 @@ materializeQueryResult(FunctionCallInfo fcinfo, (PQresultStatus(res) != PGRES_COMMAND_OK && PQresultStatus(res) != PGRES_TUPLES_OK)) { - /* - * dblink_res_error will clear the passed PGresult, so we need - * this ugly dance to avoid doing so twice during error exit - */ - PGresult *res1 = res; - - res = NULL; - dblink_res_error(conn, conname, res1, fail, + dblink_res_error(conn, conname, res, fail, "while executing query"); /* if fail isn't set, we'll return an empty query result */ } @@ -1081,7 +1059,6 @@ materializeQueryResult(FunctionCallInfo fcinfo, tuplestore_puttuple(tupstore, tuple); PQclear(res); - res = NULL; } else { @@ -1090,26 +1067,20 @@ materializeQueryResult(FunctionCallInfo fcinfo, Assert(rsinfo->setResult != NULL); PQclear(res); - res = NULL; } /* clean up data conversion short-lived memory context */ if (sinfo.tmpcontext != NULL) MemoryContextDelete(sinfo.tmpcontext); - sinfo.tmpcontext = NULL; PQclear(sinfo.last_res); - sinfo.last_res = NULL; PQclear(sinfo.cur_res); - sinfo.cur_res = NULL; } PG_CATCH(); { - /* be sure to release any libpq result we collected */ - PQclear(res); - PQclear(sinfo.last_res); - PQclear(sinfo.cur_res); - /* and clear out any pending data in libpq */ + PGresult *res; + + /* be sure to clear out any pending data in libpq */ while ((res = libpqsrv_get_result(conn, dblink_we_get_result)) != NULL) PQclear(res); @@ -1122,7 +1093,7 @@ materializeQueryResult(FunctionCallInfo fcinfo, * Execute query, and send any result rows to sinfo->tuplestore. */ static PGresult * -storeQueryResult(volatile storeInfo *sinfo, PGconn *conn, const char *sql) +storeQueryResult(storeInfo *sinfo, PGconn *conn, const char *sql) { bool first = true; int nestlevel = -1; @@ -1190,7 +1161,7 @@ storeQueryResult(volatile storeInfo *sinfo, PGconn *conn, const char *sql) * (in this case the PGresult might contain either zero or one row). */ static void -storeRow(volatile storeInfo *sinfo, PGresult *res, bool first) +storeRow(storeInfo *sinfo, PGresult *res, bool first) { int nfields = PQnfields(res); HeapTuple tuple; @@ -2795,10 +2766,13 @@ dblink_connstr_check(const char *connstr) /* * Report an error received from the remote server * - * res: the received error result (will be freed) + * res: the received error result * fail: true for ERROR ereport, false for NOTICE * fmt and following args: sprintf-style format and values for errcontext; * the resulting string should be worded like "while " + * + * If "res" is not NULL, it'll be PQclear'ed here (unless we throw error, + * in which case memory context cleanup will clear it eventually). */ static void dblink_res_error(PGconn *conn, const char *conname, PGresult *res, @@ -2806,15 +2780,11 @@ dblink_res_error(PGconn *conn, const char *conname, PGresult *res, { int level; char *pg_diag_sqlstate = PQresultErrorField(res, PG_DIAG_SQLSTATE); - char *pg_diag_message_primary = PQresultErrorField(res, PG_DIAG_MESSAGE_PRIMARY); - char *pg_diag_message_detail = PQresultErrorField(res, PG_DIAG_MESSAGE_DETAIL); - char *pg_diag_message_hint = PQresultErrorField(res, PG_DIAG_MESSAGE_HINT); - char *pg_diag_context = PQresultErrorField(res, PG_DIAG_CONTEXT); + char *message_primary = PQresultErrorField(res, PG_DIAG_MESSAGE_PRIMARY); + char *message_detail = PQresultErrorField(res, PG_DIAG_MESSAGE_DETAIL); + char *message_hint = PQresultErrorField(res, PG_DIAG_MESSAGE_HINT); + char *message_context = PQresultErrorField(res, PG_DIAG_CONTEXT); int sqlstate; - char *message_primary; - char *message_detail; - char *message_hint; - char *message_context; va_list ap; char dblink_context_msg[512]; @@ -2832,11 +2802,6 @@ dblink_res_error(PGconn *conn, const char *conname, PGresult *res, else sqlstate = ERRCODE_CONNECTION_FAILURE; - message_primary = xpstrdup(pg_diag_message_primary); - message_detail = xpstrdup(pg_diag_message_detail); - message_hint = xpstrdup(pg_diag_message_hint); - message_context = xpstrdup(pg_diag_context); - /* * If we don't get a message from the PGresult, try the PGconn. This is * needed because for connection-level failures, PQgetResult may just @@ -2845,14 +2810,6 @@ dblink_res_error(PGconn *conn, const char *conname, PGresult *res, if (message_primary == NULL) message_primary = pchomp(PQerrorMessage(conn)); - /* - * Now that we've copied all the data we need out of the PGresult, it's - * safe to free it. We must do this to avoid PGresult leakage. We're - * leaking all the strings too, but those are in palloc'd memory that will - * get cleaned up eventually. - */ - PQclear(res); - /* * Format the basic errcontext string. Below, we'll add on something * about the connection name. That's a violation of the translatability @@ -2877,6 +2834,7 @@ dblink_res_error(PGconn *conn, const char *conname, PGresult *res, dblink_context_msg, conname)) : (errcontext("%s on unnamed dblink connection", dblink_context_msg)))); + PQclear(res); } /* diff --git a/contrib/pg_stat_statements/Makefile b/contrib/pg_stat_statements/Makefile index b2bd8794d2a14..fe0478ac55266 100644 --- a/contrib/pg_stat_statements/Makefile +++ b/contrib/pg_stat_statements/Makefile @@ -7,6 +7,7 @@ OBJS = \ EXTENSION = pg_stat_statements DATA = pg_stat_statements--1.4.sql \ + pg_stat_statements--1.12--1.13.sql \ pg_stat_statements--1.11--1.12.sql pg_stat_statements--1.10--1.11.sql \ pg_stat_statements--1.9--1.10.sql pg_stat_statements--1.8--1.9.sql \ pg_stat_statements--1.7--1.8.sql pg_stat_statements--1.6--1.7.sql \ @@ -20,7 +21,7 @@ LDFLAGS_SL += $(filter -lm, $(LIBS)) REGRESS_OPTS = --temp-config $(top_srcdir)/contrib/pg_stat_statements/pg_stat_statements.conf REGRESS = select dml cursors utility level_tracking planning \ user_activity wal entry_timestamp privileges extended \ - parallel cleanup oldextversions squashing + parallel plancache cleanup oldextversions squashing # Disabled because these tests require "shared_preload_libraries=pg_stat_statements", # which typical installcheck users do not have (e.g. buildfarm clients). NO_INSTALLCHECK = 1 diff --git a/contrib/pg_stat_statements/expected/oldextversions.out b/contrib/pg_stat_statements/expected/oldextversions.out index de679b19711ab..726383a99d7c1 100644 --- a/contrib/pg_stat_statements/expected/oldextversions.out +++ b/contrib/pg_stat_statements/expected/oldextversions.out @@ -407,4 +407,71 @@ SELECT count(*) > 0 AS has_data FROM pg_stat_statements; t (1 row) +-- New functions and views for pg_stat_statements in 1.13 +AlTER EXTENSION pg_stat_statements UPDATE TO '1.13'; +\d pg_stat_statements + View "public.pg_stat_statements" + Column | Type | Collation | Nullable | Default +----------------------------+--------------------------+-----------+----------+--------- + userid | oid | | | + dbid | oid | | | + toplevel | boolean | | | + queryid | bigint | | | + query | text | | | + plans | bigint | | | + total_plan_time | double precision | | | + min_plan_time | double precision | | | + max_plan_time | double precision | | | + mean_plan_time | double precision | | | + stddev_plan_time | double precision | | | + calls | bigint | | | + total_exec_time | double precision | | | + min_exec_time | double precision | | | + max_exec_time | double precision | | | + mean_exec_time | double precision | | | + stddev_exec_time | double precision | | | + rows | bigint | | | + shared_blks_hit | bigint | | | + shared_blks_read | bigint | | | + shared_blks_dirtied | bigint | | | + shared_blks_written | bigint | | | + local_blks_hit | bigint | | | + local_blks_read | bigint | | | + local_blks_dirtied | bigint | | | + local_blks_written | bigint | | | + temp_blks_read | bigint | | | + temp_blks_written | bigint | | | + shared_blk_read_time | double precision | | | + shared_blk_write_time | double precision | | | + local_blk_read_time | double precision | | | + local_blk_write_time | double precision | | | + temp_blk_read_time | double precision | | | + temp_blk_write_time | double precision | | | + wal_records | bigint | | | + wal_fpi | bigint | | | + wal_bytes | numeric | | | + wal_buffers_full | bigint | | | + jit_functions | bigint | | | + jit_generation_time | double precision | | | + jit_inlining_count | bigint | | | + jit_inlining_time | double precision | | | + jit_optimization_count | bigint | | | + jit_optimization_time | double precision | | | + jit_emission_count | bigint | | | + jit_emission_time | double precision | | | + jit_deform_count | bigint | | | + jit_deform_time | double precision | | | + parallel_workers_to_launch | bigint | | | + parallel_workers_launched | bigint | | | + generic_plan_calls | bigint | | | + custom_plan_calls | bigint | | | + stats_since | timestamp with time zone | | | + minmax_stats_since | timestamp with time zone | | | + +SELECT count(*) > 0 AS has_data FROM pg_stat_statements; + has_data +---------- + t +(1 row) + DROP EXTENSION pg_stat_statements; diff --git a/contrib/pg_stat_statements/expected/plancache.out b/contrib/pg_stat_statements/expected/plancache.out new file mode 100644 index 0000000000000..e152de9f55130 --- /dev/null +++ b/contrib/pg_stat_statements/expected/plancache.out @@ -0,0 +1,224 @@ +-- +-- Tests with plan cache +-- +-- Setup +CREATE OR REPLACE FUNCTION select_one_func(int) RETURNS VOID AS $$ +DECLARE + ret INT; +BEGIN + SELECT $1 INTO ret; +END; +$$ LANGUAGE plpgsql; +CREATE OR REPLACE PROCEDURE select_one_proc(int) AS $$ +DECLARE + ret INT; +BEGIN + SELECT $1 INTO ret; +END; +$$ LANGUAGE plpgsql; +-- Prepared statements +SELECT pg_stat_statements_reset() IS NOT NULL AS t; + t +--- + t +(1 row) + +PREPARE p1 AS SELECT $1 AS a; +SET plan_cache_mode TO force_generic_plan; +EXECUTE p1(1); + a +--- + 1 +(1 row) + +SET plan_cache_mode TO force_custom_plan; +EXECUTE p1(1); + a +--- + 1 +(1 row) + +SELECT calls, generic_plan_calls, custom_plan_calls, query FROM pg_stat_statements + ORDER BY query COLLATE "C"; + calls | generic_plan_calls | custom_plan_calls | query +-------+--------------------+-------------------+---------------------------------------------------- + 2 | 1 | 1 | PREPARE p1 AS SELECT $1 AS a + 1 | 0 | 0 | SELECT pg_stat_statements_reset() IS NOT NULL AS t + 2 | 0 | 0 | SET plan_cache_mode TO $1 +(3 rows) + +DEALLOCATE p1; +-- Extended query protocol +SELECT pg_stat_statements_reset() IS NOT NULL AS t; + t +--- + t +(1 row) + +SELECT $1 AS a \parse p1 +SET plan_cache_mode TO force_generic_plan; +\bind_named p1 1 +; + a +--- + 1 +(1 row) + +SET plan_cache_mode TO force_custom_plan; +\bind_named p1 1 +; + a +--- + 1 +(1 row) + +SELECT calls, generic_plan_calls, custom_plan_calls, query FROM pg_stat_statements + ORDER BY query COLLATE "C"; + calls | generic_plan_calls | custom_plan_calls | query +-------+--------------------+-------------------+---------------------------------------------------- + 2 | 1 | 1 | SELECT $1 AS a + 1 | 0 | 0 | SELECT pg_stat_statements_reset() IS NOT NULL AS t + 2 | 0 | 0 | SET plan_cache_mode TO $1 +(3 rows) + +\close_prepared p1 +-- EXPLAIN [ANALYZE] EXECUTE +SET pg_stat_statements.track = 'all'; +SELECT pg_stat_statements_reset() IS NOT NULL AS t; + t +--- + t +(1 row) + +PREPARE p1 AS SELECT $1; +SET plan_cache_mode TO force_generic_plan; +EXPLAIN (COSTS OFF) EXECUTE p1(1); + QUERY PLAN +------------ + Result +(1 row) + +EXPLAIN (ANALYZE, COSTS OFF, SUMMARY OFF, TIMING OFF, BUFFERS OFF) EXECUTE p1(1); + QUERY PLAN +----------------------------------- + Result (actual rows=1.00 loops=1) +(1 row) + +SET plan_cache_mode TO force_custom_plan; +EXPLAIN (COSTS OFF) EXECUTE p1(1); + QUERY PLAN +------------ + Result +(1 row) + +EXPLAIN (ANALYZE, COSTS OFF, SUMMARY OFF, TIMING OFF, BUFFERS OFF) EXECUTE p1(1); + QUERY PLAN +----------------------------------- + Result (actual rows=1.00 loops=1) +(1 row) + +SELECT calls, generic_plan_calls, custom_plan_calls, toplevel, query FROM pg_stat_statements + ORDER BY query COLLATE "C"; + calls | generic_plan_calls | custom_plan_calls | toplevel | query +-------+--------------------+-------------------+----------+---------------------------------------------------------------------------------- + 2 | 0 | 0 | t | EXPLAIN (ANALYZE, COSTS OFF, SUMMARY OFF, TIMING OFF, BUFFERS OFF) EXECUTE p1(1) + 2 | 0 | 0 | t | EXPLAIN (COSTS OFF) EXECUTE p1(1) + 4 | 2 | 2 | f | PREPARE p1 AS SELECT $1 + 1 | 0 | 0 | t | SELECT pg_stat_statements_reset() IS NOT NULL AS t + 2 | 0 | 0 | t | SET plan_cache_mode TO $1 +(5 rows) + +RESET pg_stat_statements.track; +DEALLOCATE p1; +-- Functions/procedures +SET pg_stat_statements.track = 'all'; +SELECT pg_stat_statements_reset() IS NOT NULL AS t; + t +--- + t +(1 row) + +SET plan_cache_mode TO force_generic_plan; +SELECT select_one_func(1); + select_one_func +----------------- + +(1 row) + +CALL select_one_proc(1); +SET plan_cache_mode TO force_custom_plan; +SELECT select_one_func(1); + select_one_func +----------------- + +(1 row) + +CALL select_one_proc(1); +SELECT calls, generic_plan_calls, custom_plan_calls, toplevel, query FROM pg_stat_statements + ORDER BY query COLLATE "C"; + calls | generic_plan_calls | custom_plan_calls | toplevel | query +-------+--------------------+-------------------+----------+---------------------------------------------------- + 2 | 0 | 0 | t | CALL select_one_proc($1) + 4 | 2 | 2 | f | SELECT $1 + 1 | 0 | 0 | t | SELECT pg_stat_statements_reset() IS NOT NULL AS t + 2 | 0 | 0 | t | SELECT select_one_func($1) + 2 | 0 | 0 | t | SET plan_cache_mode TO $1 +(5 rows) + +-- +-- EXPLAIN [ANALYZE] EXECUTE + functions/procedures +-- +SET pg_stat_statements.track = 'all'; +SELECT pg_stat_statements_reset() IS NOT NULL AS t; + t +--- + t +(1 row) + +SET plan_cache_mode TO force_generic_plan; +EXPLAIN (ANALYZE, COSTS OFF, SUMMARY OFF, TIMING OFF, BUFFERS OFF) SELECT select_one_func(1); + QUERY PLAN +----------------------------------- + Result (actual rows=1.00 loops=1) +(1 row) + +EXPLAIN (COSTS OFF) SELECT select_one_func(1); + QUERY PLAN +------------ + Result +(1 row) + +CALL select_one_proc(1); +SET plan_cache_mode TO force_custom_plan; +EXPLAIN (ANALYZE, COSTS OFF, SUMMARY OFF, TIMING OFF, BUFFERS OFF) SELECT select_one_func(1); + QUERY PLAN +----------------------------------- + Result (actual rows=1.00 loops=1) +(1 row) + +EXPLAIN (COSTS OFF) SELECT select_one_func(1); + QUERY PLAN +------------ + Result +(1 row) + +CALL select_one_proc(1); +SELECT calls, generic_plan_calls, custom_plan_calls, toplevel, query FROM pg_stat_statements + ORDER BY query COLLATE "C", toplevel; + calls | generic_plan_calls | custom_plan_calls | toplevel | query +-------+--------------------+-------------------+----------+------------------------------------------------------------------------------------------------ + 2 | 0 | 0 | t | CALL select_one_proc($1) + 2 | 0 | 0 | t | EXPLAIN (ANALYZE, COSTS OFF, SUMMARY OFF, TIMING OFF, BUFFERS OFF) SELECT select_one_func($1) + 4 | 0 | 0 | f | EXPLAIN (ANALYZE, COSTS OFF, SUMMARY OFF, TIMING OFF, BUFFERS OFF) SELECT select_one_func($1); + 2 | 0 | 0 | t | EXPLAIN (COSTS OFF) SELECT select_one_func($1) + 4 | 2 | 2 | f | SELECT $1 + 1 | 0 | 0 | t | SELECT pg_stat_statements_reset() IS NOT NULL AS t + 2 | 0 | 0 | t | SET plan_cache_mode TO $1 +(7 rows) + +RESET pg_stat_statements.track; +-- +-- Cleanup +-- +DROP FUNCTION select_one_func(int); +DROP PROCEDURE select_one_proc(int); diff --git a/contrib/pg_stat_statements/meson.build b/contrib/pg_stat_statements/meson.build index 01a6cbdcf6139..7b8bfbb1de78c 100644 --- a/contrib/pg_stat_statements/meson.build +++ b/contrib/pg_stat_statements/meson.build @@ -21,6 +21,7 @@ contrib_targets += pg_stat_statements install_data( 'pg_stat_statements.control', 'pg_stat_statements--1.4.sql', + 'pg_stat_statements--1.12--1.13.sql', 'pg_stat_statements--1.11--1.12.sql', 'pg_stat_statements--1.10--1.11.sql', 'pg_stat_statements--1.9--1.10.sql', @@ -54,6 +55,7 @@ tests += { 'privileges', 'extended', 'parallel', + 'plancache', 'cleanup', 'oldextversions', 'squashing', diff --git a/contrib/pg_stat_statements/pg_stat_statements--1.12--1.13.sql b/contrib/pg_stat_statements/pg_stat_statements--1.12--1.13.sql new file mode 100644 index 0000000000000..2f0eaf14ec34d --- /dev/null +++ b/contrib/pg_stat_statements/pg_stat_statements--1.12--1.13.sql @@ -0,0 +1,78 @@ +/* contrib/pg_stat_statements/pg_stat_statements--1.12--1.13.sql */ + +-- complain if script is sourced in psql, rather than via ALTER EXTENSION +\echo Use "ALTER EXTENSION pg_stat_statements UPDATE TO '1.13'" to load this file. \quit + +/* First we have to remove them from the extension */ +ALTER EXTENSION pg_stat_statements DROP VIEW pg_stat_statements; +ALTER EXTENSION pg_stat_statements DROP FUNCTION pg_stat_statements(boolean); + +/* Then we can drop them */ +DROP VIEW pg_stat_statements; +DROP FUNCTION pg_stat_statements(boolean); + +/* Now redefine */ +CREATE FUNCTION pg_stat_statements(IN showtext boolean, + OUT userid oid, + OUT dbid oid, + OUT toplevel bool, + OUT queryid bigint, + OUT query text, + OUT plans int8, + OUT total_plan_time float8, + OUT min_plan_time float8, + OUT max_plan_time float8, + OUT mean_plan_time float8, + OUT stddev_plan_time float8, + OUT calls int8, + OUT total_exec_time float8, + OUT min_exec_time float8, + OUT max_exec_time float8, + OUT mean_exec_time float8, + OUT stddev_exec_time float8, + OUT rows int8, + OUT shared_blks_hit int8, + OUT shared_blks_read int8, + OUT shared_blks_dirtied int8, + OUT shared_blks_written int8, + OUT local_blks_hit int8, + OUT local_blks_read int8, + OUT local_blks_dirtied int8, + OUT local_blks_written int8, + OUT temp_blks_read int8, + OUT temp_blks_written int8, + OUT shared_blk_read_time float8, + OUT shared_blk_write_time float8, + OUT local_blk_read_time float8, + OUT local_blk_write_time float8, + OUT temp_blk_read_time float8, + OUT temp_blk_write_time float8, + OUT wal_records int8, + OUT wal_fpi int8, + OUT wal_bytes numeric, + OUT wal_buffers_full int8, + OUT jit_functions int8, + OUT jit_generation_time float8, + OUT jit_inlining_count int8, + OUT jit_inlining_time float8, + OUT jit_optimization_count int8, + OUT jit_optimization_time float8, + OUT jit_emission_count int8, + OUT jit_emission_time float8, + OUT jit_deform_count int8, + OUT jit_deform_time float8, + OUT parallel_workers_to_launch int8, + OUT parallel_workers_launched int8, + OUT generic_plan_calls int8, + OUT custom_plan_calls int8, + OUT stats_since timestamp with time zone, + OUT minmax_stats_since timestamp with time zone +) +RETURNS SETOF record +AS 'MODULE_PATHNAME', 'pg_stat_statements_1_13' +LANGUAGE C STRICT VOLATILE PARALLEL SAFE; + +CREATE VIEW pg_stat_statements AS + SELECT * FROM pg_stat_statements(true); + +GRANT SELECT ON pg_stat_statements TO PUBLIC; diff --git a/contrib/pg_stat_statements/pg_stat_statements.c b/contrib/pg_stat_statements/pg_stat_statements.c index e7857f81ec057..9fc9635d3300d 100644 --- a/contrib/pg_stat_statements/pg_stat_statements.c +++ b/contrib/pg_stat_statements/pg_stat_statements.c @@ -85,7 +85,7 @@ PG_MODULE_MAGIC_EXT( #define PGSS_TEXT_FILE PG_STAT_TMP_DIR "/pgss_query_texts.stat" /* Magic number identifying the stats file format */ -static const uint32 PGSS_FILE_HEADER = 0x20220408; +static const uint32 PGSS_FILE_HEADER = 0x20250731; /* PostgreSQL major version number, changes in which invalidate all entries */ static const uint32 PGSS_PG_MAJOR_VERSION = PG_VERSION_NUM / 100; @@ -114,6 +114,7 @@ typedef enum pgssVersion PGSS_V1_10, PGSS_V1_11, PGSS_V1_12, + PGSS_V1_13, } pgssVersion; typedef enum pgssStoreKind @@ -210,6 +211,8 @@ typedef struct Counters * to be launched */ int64 parallel_workers_launched; /* # of parallel workers actually * launched */ + int64 generic_plan_calls; /* number of calls using a generic plan */ + int64 custom_plan_calls; /* number of calls using a custom plan */ } Counters; /* @@ -323,6 +326,7 @@ PG_FUNCTION_INFO_V1(pg_stat_statements_1_9); PG_FUNCTION_INFO_V1(pg_stat_statements_1_10); PG_FUNCTION_INFO_V1(pg_stat_statements_1_11); PG_FUNCTION_INFO_V1(pg_stat_statements_1_12); +PG_FUNCTION_INFO_V1(pg_stat_statements_1_13); PG_FUNCTION_INFO_V1(pg_stat_statements); PG_FUNCTION_INFO_V1(pg_stat_statements_info); @@ -355,7 +359,8 @@ static void pgss_store(const char *query, int64 queryId, const struct JitInstrumentation *jitusage, JumbleState *jstate, int parallel_workers_to_launch, - int parallel_workers_launched); + int parallel_workers_launched, + PlannedStmtOrigin planOrigin); static void pg_stat_statements_internal(FunctionCallInfo fcinfo, pgssVersion api_version, bool showtext); @@ -877,7 +882,8 @@ pgss_post_parse_analyze(ParseState *pstate, Query *query, JumbleState *jstate) NULL, jstate, 0, - 0); + 0, + PLAN_STMT_UNKNOWN); } /* @@ -957,7 +963,8 @@ pgss_planner(Query *parse, NULL, NULL, 0, - 0); + 0, + result->planOrigin); } else { @@ -1091,7 +1098,8 @@ pgss_ExecutorEnd(QueryDesc *queryDesc) queryDesc->estate->es_jit ? &queryDesc->estate->es_jit->instr : NULL, NULL, queryDesc->estate->es_parallel_workers_to_launch, - queryDesc->estate->es_parallel_workers_launched); + queryDesc->estate->es_parallel_workers_launched, + queryDesc->plannedstmt->planOrigin); } if (prev_ExecutorEnd) @@ -1224,7 +1232,8 @@ pgss_ProcessUtility(PlannedStmt *pstmt, const char *queryString, NULL, NULL, 0, - 0); + 0, + pstmt->planOrigin); } else { @@ -1287,7 +1296,8 @@ pgss_store(const char *query, int64 queryId, const struct JitInstrumentation *jitusage, JumbleState *jstate, int parallel_workers_to_launch, - int parallel_workers_launched) + int parallel_workers_launched, + PlannedStmtOrigin planOrigin) { pgssHashKey key; pgssEntry *entry; @@ -1495,6 +1505,12 @@ pgss_store(const char *query, int64 queryId, entry->counters.parallel_workers_to_launch += parallel_workers_to_launch; entry->counters.parallel_workers_launched += parallel_workers_launched; + /* plan cache counters */ + if (planOrigin == PLAN_STMT_CACHE_GENERIC) + entry->counters.generic_plan_calls++; + else if (planOrigin == PLAN_STMT_CACHE_CUSTOM) + entry->counters.custom_plan_calls++; + SpinLockRelease(&entry->mutex); } @@ -1562,7 +1578,8 @@ pg_stat_statements_reset(PG_FUNCTION_ARGS) #define PG_STAT_STATEMENTS_COLS_V1_10 43 #define PG_STAT_STATEMENTS_COLS_V1_11 49 #define PG_STAT_STATEMENTS_COLS_V1_12 52 -#define PG_STAT_STATEMENTS_COLS 52 /* maximum of above */ +#define PG_STAT_STATEMENTS_COLS_V1_13 54 +#define PG_STAT_STATEMENTS_COLS 54 /* maximum of above */ /* * Retrieve statement statistics. @@ -1574,6 +1591,16 @@ pg_stat_statements_reset(PG_FUNCTION_ARGS) * expected API version is identified by embedding it in the C name of the * function. Unfortunately we weren't bright enough to do that for 1.1. */ +Datum +pg_stat_statements_1_13(PG_FUNCTION_ARGS) +{ + bool showtext = PG_GETARG_BOOL(0); + + pg_stat_statements_internal(fcinfo, PGSS_V1_13, showtext); + + return (Datum) 0; +} + Datum pg_stat_statements_1_12(PG_FUNCTION_ARGS) { @@ -1732,6 +1759,10 @@ pg_stat_statements_internal(FunctionCallInfo fcinfo, if (api_version != PGSS_V1_12) elog(ERROR, "incorrect number of output arguments"); break; + case PG_STAT_STATEMENTS_COLS_V1_13: + if (api_version != PGSS_V1_13) + elog(ERROR, "incorrect number of output arguments"); + break; default: elog(ERROR, "incorrect number of output arguments"); } @@ -1984,6 +2015,11 @@ pg_stat_statements_internal(FunctionCallInfo fcinfo, values[i++] = Int64GetDatumFast(tmp.parallel_workers_to_launch); values[i++] = Int64GetDatumFast(tmp.parallel_workers_launched); } + if (api_version >= PGSS_V1_13) + { + values[i++] = Int64GetDatumFast(tmp.generic_plan_calls); + values[i++] = Int64GetDatumFast(tmp.custom_plan_calls); + } if (api_version >= PGSS_V1_11) { values[i++] = TimestampTzGetDatum(stats_since); @@ -1999,6 +2035,7 @@ pg_stat_statements_internal(FunctionCallInfo fcinfo, api_version == PGSS_V1_10 ? PG_STAT_STATEMENTS_COLS_V1_10 : api_version == PGSS_V1_11 ? PG_STAT_STATEMENTS_COLS_V1_11 : api_version == PGSS_V1_12 ? PG_STAT_STATEMENTS_COLS_V1_12 : + api_version == PGSS_V1_13 ? PG_STAT_STATEMENTS_COLS_V1_13 : -1 /* fail if you forget to update this assert */ )); tuplestore_putvalues(rsinfo->setResult, rsinfo->setDesc, values, nulls); diff --git a/contrib/pg_stat_statements/pg_stat_statements.control b/contrib/pg_stat_statements/pg_stat_statements.control index d45ebc12e3605..2eee0ceffa894 100644 --- a/contrib/pg_stat_statements/pg_stat_statements.control +++ b/contrib/pg_stat_statements/pg_stat_statements.control @@ -1,5 +1,5 @@ # pg_stat_statements extension comment = 'track planning and execution statistics of all SQL statements executed' -default_version = '1.12' +default_version = '1.13' module_pathname = '$libdir/pg_stat_statements' relocatable = true diff --git a/contrib/pg_stat_statements/sql/oldextversions.sql b/contrib/pg_stat_statements/sql/oldextversions.sql index 13b8ca28586d1..e416efe9ffbee 100644 --- a/contrib/pg_stat_statements/sql/oldextversions.sql +++ b/contrib/pg_stat_statements/sql/oldextversions.sql @@ -63,4 +63,9 @@ AlTER EXTENSION pg_stat_statements UPDATE TO '1.12'; \d pg_stat_statements SELECT count(*) > 0 AS has_data FROM pg_stat_statements; +-- New functions and views for pg_stat_statements in 1.13 +AlTER EXTENSION pg_stat_statements UPDATE TO '1.13'; +\d pg_stat_statements +SELECT count(*) > 0 AS has_data FROM pg_stat_statements; + DROP EXTENSION pg_stat_statements; diff --git a/contrib/pg_stat_statements/sql/plancache.sql b/contrib/pg_stat_statements/sql/plancache.sql new file mode 100644 index 0000000000000..160ced7add368 --- /dev/null +++ b/contrib/pg_stat_statements/sql/plancache.sql @@ -0,0 +1,94 @@ +-- +-- Tests with plan cache +-- + +-- Setup +CREATE OR REPLACE FUNCTION select_one_func(int) RETURNS VOID AS $$ +DECLARE + ret INT; +BEGIN + SELECT $1 INTO ret; +END; +$$ LANGUAGE plpgsql; +CREATE OR REPLACE PROCEDURE select_one_proc(int) AS $$ +DECLARE + ret INT; +BEGIN + SELECT $1 INTO ret; +END; +$$ LANGUAGE plpgsql; + +-- Prepared statements +SELECT pg_stat_statements_reset() IS NOT NULL AS t; +PREPARE p1 AS SELECT $1 AS a; +SET plan_cache_mode TO force_generic_plan; +EXECUTE p1(1); +SET plan_cache_mode TO force_custom_plan; +EXECUTE p1(1); +SELECT calls, generic_plan_calls, custom_plan_calls, query FROM pg_stat_statements + ORDER BY query COLLATE "C"; +DEALLOCATE p1; + +-- Extended query protocol +SELECT pg_stat_statements_reset() IS NOT NULL AS t; +SELECT $1 AS a \parse p1 +SET plan_cache_mode TO force_generic_plan; +\bind_named p1 1 +; +SET plan_cache_mode TO force_custom_plan; +\bind_named p1 1 +; +SELECT calls, generic_plan_calls, custom_plan_calls, query FROM pg_stat_statements + ORDER BY query COLLATE "C"; +\close_prepared p1 + +-- EXPLAIN [ANALYZE] EXECUTE +SET pg_stat_statements.track = 'all'; +SELECT pg_stat_statements_reset() IS NOT NULL AS t; +PREPARE p1 AS SELECT $1; +SET plan_cache_mode TO force_generic_plan; +EXPLAIN (COSTS OFF) EXECUTE p1(1); +EXPLAIN (ANALYZE, COSTS OFF, SUMMARY OFF, TIMING OFF, BUFFERS OFF) EXECUTE p1(1); +SET plan_cache_mode TO force_custom_plan; +EXPLAIN (COSTS OFF) EXECUTE p1(1); +EXPLAIN (ANALYZE, COSTS OFF, SUMMARY OFF, TIMING OFF, BUFFERS OFF) EXECUTE p1(1); +SELECT calls, generic_plan_calls, custom_plan_calls, toplevel, query FROM pg_stat_statements + ORDER BY query COLLATE "C"; +RESET pg_stat_statements.track; +DEALLOCATE p1; + +-- Functions/procedures +SET pg_stat_statements.track = 'all'; +SELECT pg_stat_statements_reset() IS NOT NULL AS t; +SET plan_cache_mode TO force_generic_plan; +SELECT select_one_func(1); +CALL select_one_proc(1); +SET plan_cache_mode TO force_custom_plan; +SELECT select_one_func(1); +CALL select_one_proc(1); +SELECT calls, generic_plan_calls, custom_plan_calls, toplevel, query FROM pg_stat_statements + ORDER BY query COLLATE "C"; + +-- +-- EXPLAIN [ANALYZE] EXECUTE + functions/procedures +-- +SET pg_stat_statements.track = 'all'; +SELECT pg_stat_statements_reset() IS NOT NULL AS t; +SET plan_cache_mode TO force_generic_plan; +EXPLAIN (ANALYZE, COSTS OFF, SUMMARY OFF, TIMING OFF, BUFFERS OFF) SELECT select_one_func(1); +EXPLAIN (COSTS OFF) SELECT select_one_func(1); +CALL select_one_proc(1); +SET plan_cache_mode TO force_custom_plan; +EXPLAIN (ANALYZE, COSTS OFF, SUMMARY OFF, TIMING OFF, BUFFERS OFF) SELECT select_one_func(1); +EXPLAIN (COSTS OFF) SELECT select_one_func(1); +CALL select_one_proc(1); +SELECT calls, generic_plan_calls, custom_plan_calls, toplevel, query FROM pg_stat_statements + ORDER BY query COLLATE "C", toplevel; + +RESET pg_stat_statements.track; + +-- +-- Cleanup +-- +DROP FUNCTION select_one_func(int); +DROP PROCEDURE select_one_proc(int); diff --git a/contrib/postgres_fdw/connection.c b/contrib/postgres_fdw/connection.c index c1ce6f3343665..e8148f2c5a223 100644 --- a/contrib/postgres_fdw/connection.c +++ b/contrib/postgres_fdw/connection.c @@ -142,6 +142,8 @@ static void do_sql_command_begin(PGconn *conn, const char *sql); static void do_sql_command_end(PGconn *conn, const char *sql, bool consume_input); static void begin_remote_xact(ConnCacheEntry *entry); +static void pgfdw_report_internal(int elevel, PGresult *res, PGconn *conn, + const char *sql); static void pgfdw_xact_callback(XactEvent event, void *arg); static void pgfdw_subxact_callback(SubXactEvent event, SubTransactionId mySubid, @@ -815,7 +817,7 @@ static void do_sql_command_begin(PGconn *conn, const char *sql) { if (!PQsendQuery(conn, sql)) - pgfdw_report_error(ERROR, NULL, conn, false, sql); + pgfdw_report_error(NULL, conn, sql); } static void @@ -830,10 +832,10 @@ do_sql_command_end(PGconn *conn, const char *sql, bool consume_input) * would be large compared to the overhead of PQconsumeInput.) */ if (consume_input && !PQconsumeInput(conn)) - pgfdw_report_error(ERROR, NULL, conn, false, sql); + pgfdw_report_error(NULL, conn, sql); res = pgfdw_get_result(conn); if (PQresultStatus(res) != PGRES_COMMAND_OK) - pgfdw_report_error(ERROR, res, conn, true, sql); + pgfdw_report_error(res, conn, sql); PQclear(res); } @@ -966,63 +968,73 @@ pgfdw_get_result(PGconn *conn) /* * Report an error we got from the remote server. * - * elevel: error level to use (typically ERROR, but might be less) - * res: PGresult containing the error + * Callers should use pgfdw_report_error() to throw an error, or use + * pgfdw_report() for lesser message levels. (We make this distinction + * so that pgfdw_report_error() can be marked noreturn.) + * + * res: PGresult containing the error (might be NULL) * conn: connection we did the query on - * clear: if true, PQclear the result (otherwise caller will handle it) * sql: NULL, or text of remote command we tried to execute * + * If "res" is not NULL, it'll be PQclear'ed here (unless we throw error, + * in which case memory context cleanup will clear it eventually). + * * Note: callers that choose not to throw ERROR for a remote error are * responsible for making sure that the associated ConnCacheEntry gets * marked with have_error = true. */ void -pgfdw_report_error(int elevel, PGresult *res, PGconn *conn, - bool clear, const char *sql) +pgfdw_report_error(PGresult *res, PGconn *conn, const char *sql) { - /* If requested, PGresult must be released before leaving this function. */ - PG_TRY(); - { - char *diag_sqlstate = PQresultErrorField(res, PG_DIAG_SQLSTATE); - char *message_primary = PQresultErrorField(res, PG_DIAG_MESSAGE_PRIMARY); - char *message_detail = PQresultErrorField(res, PG_DIAG_MESSAGE_DETAIL); - char *message_hint = PQresultErrorField(res, PG_DIAG_MESSAGE_HINT); - char *message_context = PQresultErrorField(res, PG_DIAG_CONTEXT); - int sqlstate; - - if (diag_sqlstate) - sqlstate = MAKE_SQLSTATE(diag_sqlstate[0], - diag_sqlstate[1], - diag_sqlstate[2], - diag_sqlstate[3], - diag_sqlstate[4]); - else - sqlstate = ERRCODE_CONNECTION_FAILURE; + pgfdw_report_internal(ERROR, res, conn, sql); + pg_unreachable(); +} - /* - * If we don't get a message from the PGresult, try the PGconn. This - * is needed because for connection-level failures, PQgetResult may - * just return NULL, not a PGresult at all. - */ - if (message_primary == NULL) - message_primary = pchomp(PQerrorMessage(conn)); - - ereport(elevel, - (errcode(sqlstate), - (message_primary != NULL && message_primary[0] != '\0') ? - errmsg_internal("%s", message_primary) : - errmsg("could not obtain message string for remote error"), - message_detail ? errdetail_internal("%s", message_detail) : 0, - message_hint ? errhint("%s", message_hint) : 0, - message_context ? errcontext("%s", message_context) : 0, - sql ? errcontext("remote SQL command: %s", sql) : 0)); - } - PG_FINALLY(); - { - if (clear) - PQclear(res); - } - PG_END_TRY(); +void +pgfdw_report(int elevel, PGresult *res, PGconn *conn, const char *sql) +{ + Assert(elevel < ERROR); /* use pgfdw_report_error for that */ + pgfdw_report_internal(elevel, res, conn, sql); +} + +static void +pgfdw_report_internal(int elevel, PGresult *res, PGconn *conn, + const char *sql) +{ + char *diag_sqlstate = PQresultErrorField(res, PG_DIAG_SQLSTATE); + char *message_primary = PQresultErrorField(res, PG_DIAG_MESSAGE_PRIMARY); + char *message_detail = PQresultErrorField(res, PG_DIAG_MESSAGE_DETAIL); + char *message_hint = PQresultErrorField(res, PG_DIAG_MESSAGE_HINT); + char *message_context = PQresultErrorField(res, PG_DIAG_CONTEXT); + int sqlstate; + + if (diag_sqlstate) + sqlstate = MAKE_SQLSTATE(diag_sqlstate[0], + diag_sqlstate[1], + diag_sqlstate[2], + diag_sqlstate[3], + diag_sqlstate[4]); + else + sqlstate = ERRCODE_CONNECTION_FAILURE; + + /* + * If we don't get a message from the PGresult, try the PGconn. This is + * needed because for connection-level failures, PQgetResult may just + * return NULL, not a PGresult at all. + */ + if (message_primary == NULL) + message_primary = pchomp(PQerrorMessage(conn)); + + ereport(elevel, + (errcode(sqlstate), + (message_primary != NULL && message_primary[0] != '\0') ? + errmsg_internal("%s", message_primary) : + errmsg("could not obtain message string for remote error"), + message_detail ? errdetail_internal("%s", message_detail) : 0, + message_hint ? errhint("%s", message_hint) : 0, + message_context ? errcontext("%s", message_context) : 0, + sql ? errcontext("remote SQL command: %s", sql) : 0)); + PQclear(res); } /* @@ -1545,7 +1557,7 @@ pgfdw_exec_cleanup_query_begin(PGconn *conn, const char *query) */ if (!PQsendQuery(conn, query)) { - pgfdw_report_error(WARNING, NULL, conn, false, query); + pgfdw_report(WARNING, NULL, conn, query); return false; } @@ -1570,7 +1582,7 @@ pgfdw_exec_cleanup_query_end(PGconn *conn, const char *query, */ if (consume_input && !PQconsumeInput(conn)) { - pgfdw_report_error(WARNING, NULL, conn, false, query); + pgfdw_report(WARNING, NULL, conn, query); return false; } @@ -1582,7 +1594,7 @@ pgfdw_exec_cleanup_query_end(PGconn *conn, const char *query, (errmsg("could not get query result due to timeout"), errcontext("remote SQL command: %s", query))); else - pgfdw_report_error(WARNING, NULL, conn, false, query); + pgfdw_report(WARNING, NULL, conn, query); return false; } @@ -1590,7 +1602,7 @@ pgfdw_exec_cleanup_query_end(PGconn *conn, const char *query, /* Issue a warning if not successful. */ if (PQresultStatus(result) != PGRES_COMMAND_OK) { - pgfdw_report_error(WARNING, result, conn, true, query); + pgfdw_report(WARNING, result, conn, query); return ignore_errors; } PQclear(result); @@ -1618,103 +1630,90 @@ pgfdw_get_cleanup_result(PGconn *conn, TimestampTz endtime, PGresult **result, bool *timed_out) { - volatile bool failed = false; - PGresult *volatile last_res = NULL; + bool failed = false; + PGresult *last_res = NULL; + int canceldelta = RETRY_CANCEL_TIMEOUT * 2; *result = NULL; *timed_out = false; - - /* In what follows, do not leak any PGresults on an error. */ - PG_TRY(); + for (;;) { - int canceldelta = RETRY_CANCEL_TIMEOUT * 2; + PGresult *res; - for (;;) + while (PQisBusy(conn)) { - PGresult *res; + int wc; + TimestampTz now = GetCurrentTimestamp(); + long cur_timeout; - while (PQisBusy(conn)) + /* If timeout has expired, give up. */ + if (now >= endtime) { - int wc; - TimestampTz now = GetCurrentTimestamp(); - long cur_timeout; - - /* If timeout has expired, give up. */ - if (now >= endtime) - { - *timed_out = true; - failed = true; - goto exit; - } + *timed_out = true; + failed = true; + goto exit; + } - /* If we need to re-issue the cancel request, do that. */ - if (now >= retrycanceltime) - { - /* We ignore failure to issue the repeated request. */ - (void) libpqsrv_cancel(conn, endtime); + /* If we need to re-issue the cancel request, do that. */ + if (now >= retrycanceltime) + { + /* We ignore failure to issue the repeated request. */ + (void) libpqsrv_cancel(conn, endtime); - /* Recompute "now" in case that took measurable time. */ - now = GetCurrentTimestamp(); + /* Recompute "now" in case that took measurable time. */ + now = GetCurrentTimestamp(); - /* Adjust re-cancel timeout in increasing steps. */ - retrycanceltime = TimestampTzPlusMilliseconds(now, - canceldelta); - canceldelta += canceldelta; - } + /* Adjust re-cancel timeout in increasing steps. */ + retrycanceltime = TimestampTzPlusMilliseconds(now, + canceldelta); + canceldelta += canceldelta; + } - /* If timeout has expired, give up, else get sleep time. */ - cur_timeout = TimestampDifferenceMilliseconds(now, - Min(endtime, - retrycanceltime)); - if (cur_timeout <= 0) - { - *timed_out = true; - failed = true; - goto exit; - } + /* If timeout has expired, give up, else get sleep time. */ + cur_timeout = TimestampDifferenceMilliseconds(now, + Min(endtime, + retrycanceltime)); + if (cur_timeout <= 0) + { + *timed_out = true; + failed = true; + goto exit; + } - /* first time, allocate or get the custom wait event */ - if (pgfdw_we_cleanup_result == 0) - pgfdw_we_cleanup_result = WaitEventExtensionNew("PostgresFdwCleanupResult"); + /* first time, allocate or get the custom wait event */ + if (pgfdw_we_cleanup_result == 0) + pgfdw_we_cleanup_result = WaitEventExtensionNew("PostgresFdwCleanupResult"); - /* Sleep until there's something to do */ - wc = WaitLatchOrSocket(MyLatch, - WL_LATCH_SET | WL_SOCKET_READABLE | - WL_TIMEOUT | WL_EXIT_ON_PM_DEATH, - PQsocket(conn), - cur_timeout, pgfdw_we_cleanup_result); - ResetLatch(MyLatch); + /* Sleep until there's something to do */ + wc = WaitLatchOrSocket(MyLatch, + WL_LATCH_SET | WL_SOCKET_READABLE | + WL_TIMEOUT | WL_EXIT_ON_PM_DEATH, + PQsocket(conn), + cur_timeout, pgfdw_we_cleanup_result); + ResetLatch(MyLatch); - CHECK_FOR_INTERRUPTS(); + CHECK_FOR_INTERRUPTS(); - /* Data available in socket? */ - if (wc & WL_SOCKET_READABLE) + /* Data available in socket? */ + if (wc & WL_SOCKET_READABLE) + { + if (!PQconsumeInput(conn)) { - if (!PQconsumeInput(conn)) - { - /* connection trouble */ - failed = true; - goto exit; - } + /* connection trouble */ + failed = true; + goto exit; } } + } - res = PQgetResult(conn); - if (res == NULL) - break; /* query is complete */ + res = PQgetResult(conn); + if (res == NULL) + break; /* query is complete */ - PQclear(last_res); - last_res = res; - } -exit: ; - } - PG_CATCH(); - { PQclear(last_res); - PG_RE_THROW(); + last_res = res; } - PG_END_TRY(); - +exit: if (failed) PQclear(last_res); else diff --git a/contrib/postgres_fdw/option.c b/contrib/postgres_fdw/option.c index c2f936640bca8..d6fa89bad9399 100644 --- a/contrib/postgres_fdw/option.c +++ b/contrib/postgres_fdw/option.c @@ -21,6 +21,7 @@ #include "libpq/libpq-be.h" #include "postgres_fdw.h" #include "utils/guc.h" +#include "utils/memutils.h" #include "utils/varlena.h" /* @@ -39,12 +40,6 @@ typedef struct PgFdwOption */ static PgFdwOption *postgres_fdw_options; -/* - * Valid options for libpq. - * Allocated and filled in InitPgFdwOptions. - */ -static PQconninfoOption *libpq_options; - /* * GUC parameters */ @@ -239,6 +234,7 @@ static void InitPgFdwOptions(void) { int num_libpq_opts; + PQconninfoOption *libpq_options; PQconninfoOption *lopt; PgFdwOption *popt; @@ -307,8 +303,8 @@ InitPgFdwOptions(void) * Get list of valid libpq options. * * To avoid unnecessary work, we get the list once and use it throughout - * the lifetime of this backend process. We don't need to care about - * memory context issues, because PQconndefaults allocates with malloc. + * the lifetime of this backend process. Hence, we'll allocate it in + * TopMemoryContext. */ libpq_options = PQconndefaults(); if (!libpq_options) /* assume reason for failure is OOM */ @@ -325,19 +321,11 @@ InitPgFdwOptions(void) /* * Construct an array which consists of all valid options for * postgres_fdw, by appending FDW-specific options to libpq options. - * - * We use plain malloc here to allocate postgres_fdw_options because it - * lives as long as the backend process does. Besides, keeping - * libpq_options in memory allows us to avoid copying every keyword - * string. */ postgres_fdw_options = (PgFdwOption *) - malloc(sizeof(PgFdwOption) * num_libpq_opts + - sizeof(non_libpq_options)); - if (postgres_fdw_options == NULL) - ereport(ERROR, - (errcode(ERRCODE_FDW_OUT_OF_MEMORY), - errmsg("out of memory"))); + MemoryContextAlloc(TopMemoryContext, + sizeof(PgFdwOption) * num_libpq_opts + + sizeof(non_libpq_options)); popt = postgres_fdw_options; for (lopt = libpq_options; lopt->keyword; lopt++) @@ -355,8 +343,8 @@ InitPgFdwOptions(void) if (strncmp(lopt->keyword, "oauth_", strlen("oauth_")) == 0) continue; - /* We don't have to copy keyword string, as described above. */ - popt->keyword = lopt->keyword; + popt->keyword = MemoryContextStrdup(TopMemoryContext, + lopt->keyword); /* * "user" and any secret options are allowed only on user mappings. @@ -371,6 +359,9 @@ InitPgFdwOptions(void) popt++; } + /* Done with libpq's output structure. */ + PQconninfoFree(libpq_options); + /* Append FDW-specific options and dummy terminator. */ memcpy(popt, non_libpq_options, sizeof(non_libpq_options)); } diff --git a/contrib/postgres_fdw/postgres_fdw.c b/contrib/postgres_fdw/postgres_fdw.c index e0a34b27c7cfd..456b267f70b5b 100644 --- a/contrib/postgres_fdw/postgres_fdw.c +++ b/contrib/postgres_fdw/postgres_fdw.c @@ -240,7 +240,6 @@ typedef struct PgFdwDirectModifyState PGresult *result; /* result for query */ int num_tuples; /* # of result tuples */ int next_tuple; /* index of next one to return */ - MemoryContextCallback result_cb; /* ensures result will get freed */ Relation resultRel; /* relcache entry for the target relation */ AttrNumber *attnoMap; /* array of attnums of input user columns */ AttrNumber ctidAttno; /* attnum of input ctid column */ @@ -1703,13 +1702,9 @@ postgresReScanForeignScan(ForeignScanState *node) return; } - /* - * We don't use a PG_TRY block here, so be careful not to throw error - * without releasing the PGresult. - */ res = pgfdw_exec_query(fsstate->conn, sql, fsstate->conn_state); if (PQresultStatus(res) != PGRES_COMMAND_OK) - pgfdw_report_error(ERROR, res, fsstate->conn, true, sql); + pgfdw_report_error(res, fsstate->conn, sql); PQclear(res); /* Now force a fresh FETCH. */ @@ -2671,17 +2666,6 @@ postgresBeginDirectModify(ForeignScanState *node, int eflags) dmstate = (PgFdwDirectModifyState *) palloc0(sizeof(PgFdwDirectModifyState)); node->fdw_state = dmstate; - /* - * We use a memory context callback to ensure that the dmstate's PGresult - * (if any) will be released, even if the query fails somewhere that's - * outside our control. The callback is always armed for the duration of - * the query; this relies on PQclear(NULL) being a no-op. - */ - dmstate->result_cb.func = (MemoryContextCallbackFunction) PQclear; - dmstate->result_cb.arg = NULL; - MemoryContextRegisterResetCallback(CurrentMemoryContext, - &dmstate->result_cb); - /* * Identify which user to do the remote access as. This should match what * ExecCheckPermissions() does. @@ -2829,13 +2813,7 @@ postgresEndDirectModify(ForeignScanState *node) return; /* Release PGresult */ - if (dmstate->result) - { - PQclear(dmstate->result); - dmstate->result = NULL; - /* ... and don't forget to disable the callback */ - dmstate->result_cb.arg = NULL; - } + PQclear(dmstate->result); /* Release remote connection */ ReleaseConnection(dmstate->conn); @@ -3626,41 +3604,32 @@ get_remote_estimate(const char *sql, PGconn *conn, double *rows, int *width, Cost *startup_cost, Cost *total_cost) { - PGresult *volatile res = NULL; - - /* PGresult must be released before leaving this function. */ - PG_TRY(); - { - char *line; - char *p; - int n; + PGresult *res; + char *line; + char *p; + int n; - /* - * Execute EXPLAIN remotely. - */ - res = pgfdw_exec_query(conn, sql, NULL); - if (PQresultStatus(res) != PGRES_TUPLES_OK) - pgfdw_report_error(ERROR, res, conn, false, sql); + /* + * Execute EXPLAIN remotely. + */ + res = pgfdw_exec_query(conn, sql, NULL); + if (PQresultStatus(res) != PGRES_TUPLES_OK) + pgfdw_report_error(res, conn, sql); - /* - * Extract cost numbers for topmost plan node. Note we search for a - * left paren from the end of the line to avoid being confused by - * other uses of parentheses. - */ - line = PQgetvalue(res, 0, 0); - p = strrchr(line, '('); - if (p == NULL) - elog(ERROR, "could not interpret EXPLAIN output: \"%s\"", line); - n = sscanf(p, "(cost=%lf..%lf rows=%lf width=%d)", - startup_cost, total_cost, rows, width); - if (n != 4) - elog(ERROR, "could not interpret EXPLAIN output: \"%s\"", line); - } - PG_FINALLY(); - { - PQclear(res); - } - PG_END_TRY(); + /* + * Extract cost numbers for topmost plan node. Note we search for a left + * paren from the end of the line to avoid being confused by other uses of + * parentheses. + */ + line = PQgetvalue(res, 0, 0); + p = strrchr(line, '('); + if (p == NULL) + elog(ERROR, "could not interpret EXPLAIN output: \"%s\"", line); + n = sscanf(p, "(cost=%lf..%lf rows=%lf width=%d)", + startup_cost, total_cost, rows, width); + if (n != 4) + elog(ERROR, "could not interpret EXPLAIN output: \"%s\"", line); + PQclear(res); } /* @@ -3800,17 +3769,14 @@ create_cursor(ForeignScanState *node) */ if (!PQsendQueryParams(conn, buf.data, numParams, NULL, values, NULL, NULL, 0)) - pgfdw_report_error(ERROR, NULL, conn, false, buf.data); + pgfdw_report_error(NULL, conn, buf.data); /* * Get the result, and check for success. - * - * We don't use a PG_TRY block here, so be careful not to throw error - * without releasing the PGresult. */ res = pgfdw_get_result(conn); if (PQresultStatus(res) != PGRES_COMMAND_OK) - pgfdw_report_error(ERROR, res, conn, true, fsstate->query); + pgfdw_report_error(res, conn, fsstate->query); PQclear(res); /* Mark the cursor as created, and show no tuples have been retrieved */ @@ -3832,7 +3798,10 @@ static void fetch_more_data(ForeignScanState *node) { PgFdwScanState *fsstate = (PgFdwScanState *) node->fdw_state; - PGresult *volatile res = NULL; + PGconn *conn = fsstate->conn; + PGresult *res; + int numrows; + int i; MemoryContext oldcontext; /* @@ -3843,74 +3812,63 @@ fetch_more_data(ForeignScanState *node) MemoryContextReset(fsstate->batch_cxt); oldcontext = MemoryContextSwitchTo(fsstate->batch_cxt); - /* PGresult must be released before leaving this function. */ - PG_TRY(); + if (fsstate->async_capable) { - PGconn *conn = fsstate->conn; - int numrows; - int i; + Assert(fsstate->conn_state->pendingAreq); - if (fsstate->async_capable) - { - Assert(fsstate->conn_state->pendingAreq); + /* + * The query was already sent by an earlier call to + * fetch_more_data_begin. So now we just fetch the result. + */ + res = pgfdw_get_result(conn); + /* On error, report the original query, not the FETCH. */ + if (PQresultStatus(res) != PGRES_TUPLES_OK) + pgfdw_report_error(res, conn, fsstate->query); - /* - * The query was already sent by an earlier call to - * fetch_more_data_begin. So now we just fetch the result. - */ - res = pgfdw_get_result(conn); - /* On error, report the original query, not the FETCH. */ - if (PQresultStatus(res) != PGRES_TUPLES_OK) - pgfdw_report_error(ERROR, res, conn, false, fsstate->query); + /* Reset per-connection state */ + fsstate->conn_state->pendingAreq = NULL; + } + else + { + char sql[64]; - /* Reset per-connection state */ - fsstate->conn_state->pendingAreq = NULL; - } - else - { - char sql[64]; + /* This is a regular synchronous fetch. */ + snprintf(sql, sizeof(sql), "FETCH %d FROM c%u", + fsstate->fetch_size, fsstate->cursor_number); - /* This is a regular synchronous fetch. */ - snprintf(sql, sizeof(sql), "FETCH %d FROM c%u", - fsstate->fetch_size, fsstate->cursor_number); + res = pgfdw_exec_query(conn, sql, fsstate->conn_state); + /* On error, report the original query, not the FETCH. */ + if (PQresultStatus(res) != PGRES_TUPLES_OK) + pgfdw_report_error(res, conn, fsstate->query); + } - res = pgfdw_exec_query(conn, sql, fsstate->conn_state); - /* On error, report the original query, not the FETCH. */ - if (PQresultStatus(res) != PGRES_TUPLES_OK) - pgfdw_report_error(ERROR, res, conn, false, fsstate->query); - } + /* Convert the data into HeapTuples */ + numrows = PQntuples(res); + fsstate->tuples = (HeapTuple *) palloc0(numrows * sizeof(HeapTuple)); + fsstate->num_tuples = numrows; + fsstate->next_tuple = 0; - /* Convert the data into HeapTuples */ - numrows = PQntuples(res); - fsstate->tuples = (HeapTuple *) palloc0(numrows * sizeof(HeapTuple)); - fsstate->num_tuples = numrows; - fsstate->next_tuple = 0; + for (i = 0; i < numrows; i++) + { + Assert(IsA(node->ss.ps.plan, ForeignScan)); - for (i = 0; i < numrows; i++) - { - Assert(IsA(node->ss.ps.plan, ForeignScan)); - - fsstate->tuples[i] = - make_tuple_from_result_row(res, i, - fsstate->rel, - fsstate->attinmeta, - fsstate->retrieved_attrs, - node, - fsstate->temp_cxt); - } + fsstate->tuples[i] = + make_tuple_from_result_row(res, i, + fsstate->rel, + fsstate->attinmeta, + fsstate->retrieved_attrs, + node, + fsstate->temp_cxt); + } - /* Update fetch_ct_2 */ - if (fsstate->fetch_ct_2 < 2) - fsstate->fetch_ct_2++; + /* Update fetch_ct_2 */ + if (fsstate->fetch_ct_2 < 2) + fsstate->fetch_ct_2++; - /* Must be EOF if we didn't get as many tuples as we asked for. */ - fsstate->eof_reached = (numrows < fsstate->fetch_size); - } - PG_FINALLY(); - { - PQclear(res); - } - PG_END_TRY(); + /* Must be EOF if we didn't get as many tuples as we asked for. */ + fsstate->eof_reached = (numrows < fsstate->fetch_size); + + PQclear(res); MemoryContextSwitchTo(oldcontext); } @@ -3984,14 +3942,9 @@ close_cursor(PGconn *conn, unsigned int cursor_number, PGresult *res; snprintf(sql, sizeof(sql), "CLOSE c%u", cursor_number); - - /* - * We don't use a PG_TRY block here, so be careful not to throw error - * without releasing the PGresult. - */ res = pgfdw_exec_query(conn, sql, conn_state); if (PQresultStatus(res) != PGRES_COMMAND_OK) - pgfdw_report_error(ERROR, res, conn, true, sql); + pgfdw_report_error(res, conn, sql); PQclear(res); } @@ -4199,18 +4152,15 @@ execute_foreign_modify(EState *estate, NULL, NULL, 0)) - pgfdw_report_error(ERROR, NULL, fmstate->conn, false, fmstate->query); + pgfdw_report_error(NULL, fmstate->conn, fmstate->query); /* * Get the result, and check for success. - * - * We don't use a PG_TRY block here, so be careful not to throw error - * without releasing the PGresult. */ res = pgfdw_get_result(fmstate->conn); if (PQresultStatus(res) != (fmstate->has_returning ? PGRES_TUPLES_OK : PGRES_COMMAND_OK)) - pgfdw_report_error(ERROR, res, fmstate->conn, true, fmstate->query); + pgfdw_report_error(res, fmstate->conn, fmstate->query); /* Check number of rows affected, and fetch RETURNING tuple if any */ if (fmstate->has_returning) @@ -4269,17 +4219,14 @@ prepare_foreign_modify(PgFdwModifyState *fmstate) fmstate->query, 0, NULL)) - pgfdw_report_error(ERROR, NULL, fmstate->conn, false, fmstate->query); + pgfdw_report_error(NULL, fmstate->conn, fmstate->query); /* * Get the result, and check for success. - * - * We don't use a PG_TRY block here, so be careful not to throw error - * without releasing the PGresult. */ res = pgfdw_get_result(fmstate->conn); if (PQresultStatus(res) != PGRES_COMMAND_OK) - pgfdw_report_error(ERROR, res, fmstate->conn, true, fmstate->query); + pgfdw_report_error(res, fmstate->conn, fmstate->query); PQclear(res); /* This action shows that the prepare has been done. */ @@ -4370,37 +4317,25 @@ convert_prep_stmt_params(PgFdwModifyState *fmstate, /* * store_returning_result * Store the result of a RETURNING clause - * - * On error, be sure to release the PGresult on the way out. Callers do not - * have PG_TRY blocks to ensure this happens. */ static void store_returning_result(PgFdwModifyState *fmstate, TupleTableSlot *slot, PGresult *res) { - PG_TRY(); - { - HeapTuple newtup; + HeapTuple newtup; - newtup = make_tuple_from_result_row(res, 0, - fmstate->rel, - fmstate->attinmeta, - fmstate->retrieved_attrs, - NULL, - fmstate->temp_cxt); + newtup = make_tuple_from_result_row(res, 0, + fmstate->rel, + fmstate->attinmeta, + fmstate->retrieved_attrs, + NULL, + fmstate->temp_cxt); - /* - * The returning slot will not necessarily be suitable to store - * heaptuples directly, so allow for conversion. - */ - ExecForceStoreHeapTuple(newtup, slot, true); - } - PG_CATCH(); - { - PQclear(res); - PG_RE_THROW(); - } - PG_END_TRY(); + /* + * The returning slot will not necessarily be suitable to store heaptuples + * directly, so allow for conversion. + */ + ExecForceStoreHeapTuple(newtup, slot, true); } /* @@ -4436,14 +4371,9 @@ deallocate_query(PgFdwModifyState *fmstate) return; snprintf(sql, sizeof(sql), "DEALLOCATE %s", fmstate->p_name); - - /* - * We don't use a PG_TRY block here, so be careful not to throw error - * without releasing the PGresult. - */ res = pgfdw_exec_query(fmstate->conn, sql, fmstate->conn_state); if (PQresultStatus(res) != PGRES_COMMAND_OK) - pgfdw_report_error(ERROR, res, fmstate->conn, true, sql); + pgfdw_report_error(res, fmstate->conn, sql); PQclear(res); pfree(fmstate->p_name); fmstate->p_name = NULL; @@ -4611,24 +4541,24 @@ execute_dml_stmt(ForeignScanState *node) */ if (!PQsendQueryParams(dmstate->conn, dmstate->query, numParams, NULL, values, NULL, NULL, 0)) - pgfdw_report_error(ERROR, NULL, dmstate->conn, false, dmstate->query); + pgfdw_report_error(NULL, dmstate->conn, dmstate->query); /* * Get the result, and check for success. - * - * We use a memory context callback to ensure that the PGresult will be - * released, even if the query fails somewhere that's outside our control. - * The callback is already registered, just need to fill in its arg. */ - Assert(dmstate->result == NULL); dmstate->result = pgfdw_get_result(dmstate->conn); - dmstate->result_cb.arg = dmstate->result; - if (PQresultStatus(dmstate->result) != (dmstate->has_returning ? PGRES_TUPLES_OK : PGRES_COMMAND_OK)) - pgfdw_report_error(ERROR, dmstate->result, dmstate->conn, false, + pgfdw_report_error(dmstate->result, dmstate->conn, dmstate->query); + /* + * The result potentially needs to survive across multiple executor row + * cycles, so move it to the context where the dmstate is. + */ + dmstate->result = libpqsrv_PGresultSetParent(dmstate->result, + GetMemoryChunkContext(dmstate)); + /* Get the number of rows affected. */ if (dmstate->has_returning) dmstate->num_tuples = PQntuples(dmstate->result); @@ -4965,7 +4895,7 @@ postgresAnalyzeForeignTable(Relation relation, UserMapping *user; PGconn *conn; StringInfoData sql; - PGresult *volatile res = NULL; + PGresult *res; /* Return the row-analysis function pointer */ *func = postgresAcquireSampleRowsFunc; @@ -4991,22 +4921,14 @@ postgresAnalyzeForeignTable(Relation relation, initStringInfo(&sql); deparseAnalyzeSizeSql(&sql, relation); - /* In what follows, do not risk leaking any PGresults. */ - PG_TRY(); - { - res = pgfdw_exec_query(conn, sql.data, NULL); - if (PQresultStatus(res) != PGRES_TUPLES_OK) - pgfdw_report_error(ERROR, res, conn, false, sql.data); + res = pgfdw_exec_query(conn, sql.data, NULL); + if (PQresultStatus(res) != PGRES_TUPLES_OK) + pgfdw_report_error(res, conn, sql.data); - if (PQntuples(res) != 1 || PQnfields(res) != 1) - elog(ERROR, "unexpected result from deparseAnalyzeSizeSql query"); - *totalpages = strtoul(PQgetvalue(res, 0, 0), NULL, 10); - } - PG_FINALLY(); - { - PQclear(res); - } - PG_END_TRY(); + if (PQntuples(res) != 1 || PQnfields(res) != 1) + elog(ERROR, "unexpected result from deparseAnalyzeSizeSql query"); + *totalpages = strtoul(PQgetvalue(res, 0, 0), NULL, 10); + PQclear(res); ReleaseConnection(conn); @@ -5027,9 +4949,9 @@ postgresGetAnalyzeInfoForForeignTable(Relation relation, bool *can_tablesample) UserMapping *user; PGconn *conn; StringInfoData sql; - PGresult *volatile res = NULL; - volatile double reltuples = -1; - volatile char relkind = 0; + PGresult *res; + double reltuples; + char relkind; /* assume the remote relation does not support TABLESAMPLE */ *can_tablesample = false; @@ -5048,24 +4970,15 @@ postgresGetAnalyzeInfoForForeignTable(Relation relation, bool *can_tablesample) initStringInfo(&sql); deparseAnalyzeInfoSql(&sql, relation); - /* In what follows, do not risk leaking any PGresults. */ - PG_TRY(); - { - res = pgfdw_exec_query(conn, sql.data, NULL); - if (PQresultStatus(res) != PGRES_TUPLES_OK) - pgfdw_report_error(ERROR, res, conn, false, sql.data); + res = pgfdw_exec_query(conn, sql.data, NULL); + if (PQresultStatus(res) != PGRES_TUPLES_OK) + pgfdw_report_error(res, conn, sql.data); - if (PQntuples(res) != 1 || PQnfields(res) != 2) - elog(ERROR, "unexpected result from deparseAnalyzeInfoSql query"); - reltuples = strtod(PQgetvalue(res, 0, 0), NULL); - relkind = *(PQgetvalue(res, 0, 1)); - } - PG_FINALLY(); - { - if (res) - PQclear(res); - } - PG_END_TRY(); + if (PQntuples(res) != 1 || PQnfields(res) != 2) + elog(ERROR, "unexpected result from deparseAnalyzeInfoSql query"); + reltuples = strtod(PQgetvalue(res, 0, 0), NULL); + relkind = *(PQgetvalue(res, 0, 1)); + PQclear(res); ReleaseConnection(conn); @@ -5105,10 +5018,12 @@ postgresAcquireSampleRowsFunc(Relation relation, int elevel, int server_version_num; PgFdwSamplingMethod method = ANALYZE_SAMPLE_AUTO; /* auto is default */ double sample_frac = -1.0; - double reltuples; + double reltuples = -1.0; unsigned int cursor_number; StringInfoData sql; - PGresult *volatile res = NULL; + PGresult *res; + char fetch_sql[64]; + int fetch_size; ListCell *lc; /* Initialize workspace state */ @@ -5285,91 +5200,76 @@ postgresAcquireSampleRowsFunc(Relation relation, int elevel, deparseAnalyzeSql(&sql, relation, method, sample_frac, &astate.retrieved_attrs); - /* In what follows, do not risk leaking any PGresults. */ - PG_TRY(); - { - char fetch_sql[64]; - int fetch_size; - - res = pgfdw_exec_query(conn, sql.data, NULL); - if (PQresultStatus(res) != PGRES_COMMAND_OK) - pgfdw_report_error(ERROR, res, conn, false, sql.data); - PQclear(res); - res = NULL; + res = pgfdw_exec_query(conn, sql.data, NULL); + if (PQresultStatus(res) != PGRES_COMMAND_OK) + pgfdw_report_error(res, conn, sql.data); + PQclear(res); - /* - * Determine the fetch size. The default is arbitrary, but shouldn't - * be enormous. - */ - fetch_size = 100; - foreach(lc, server->options) - { - DefElem *def = (DefElem *) lfirst(lc); + /* + * Determine the fetch size. The default is arbitrary, but shouldn't be + * enormous. + */ + fetch_size = 100; + foreach(lc, server->options) + { + DefElem *def = (DefElem *) lfirst(lc); - if (strcmp(def->defname, "fetch_size") == 0) - { - (void) parse_int(defGetString(def), &fetch_size, 0, NULL); - break; - } - } - foreach(lc, table->options) + if (strcmp(def->defname, "fetch_size") == 0) { - DefElem *def = (DefElem *) lfirst(lc); - - if (strcmp(def->defname, "fetch_size") == 0) - { - (void) parse_int(defGetString(def), &fetch_size, 0, NULL); - break; - } + (void) parse_int(defGetString(def), &fetch_size, 0, NULL); + break; } + } + foreach(lc, table->options) + { + DefElem *def = (DefElem *) lfirst(lc); - /* Construct command to fetch rows from remote. */ - snprintf(fetch_sql, sizeof(fetch_sql), "FETCH %d FROM c%u", - fetch_size, cursor_number); - - /* Retrieve and process rows a batch at a time. */ - for (;;) + if (strcmp(def->defname, "fetch_size") == 0) { - int numrows; - int i; + (void) parse_int(defGetString(def), &fetch_size, 0, NULL); + break; + } + } - /* Allow users to cancel long query */ - CHECK_FOR_INTERRUPTS(); + /* Construct command to fetch rows from remote. */ + snprintf(fetch_sql, sizeof(fetch_sql), "FETCH %d FROM c%u", + fetch_size, cursor_number); - /* - * XXX possible future improvement: if rowstoskip is large, we - * could issue a MOVE rather than physically fetching the rows, - * then just adjust rowstoskip and samplerows appropriately. - */ + /* Retrieve and process rows a batch at a time. */ + for (;;) + { + int numrows; + int i; - /* Fetch some rows */ - res = pgfdw_exec_query(conn, fetch_sql, NULL); - /* On error, report the original query, not the FETCH. */ - if (PQresultStatus(res) != PGRES_TUPLES_OK) - pgfdw_report_error(ERROR, res, conn, false, sql.data); + /* Allow users to cancel long query */ + CHECK_FOR_INTERRUPTS(); - /* Process whatever we got. */ - numrows = PQntuples(res); - for (i = 0; i < numrows; i++) - analyze_row_processor(res, i, &astate); + /* + * XXX possible future improvement: if rowstoskip is large, we could + * issue a MOVE rather than physically fetching the rows, then just + * adjust rowstoskip and samplerows appropriately. + */ - PQclear(res); - res = NULL; + /* Fetch some rows */ + res = pgfdw_exec_query(conn, fetch_sql, NULL); + /* On error, report the original query, not the FETCH. */ + if (PQresultStatus(res) != PGRES_TUPLES_OK) + pgfdw_report_error(res, conn, sql.data); - /* Must be EOF if we didn't get all the rows requested. */ - if (numrows < fetch_size) - break; - } + /* Process whatever we got. */ + numrows = PQntuples(res); + for (i = 0; i < numrows; i++) + analyze_row_processor(res, i, &astate); - /* Close the cursor, just to be tidy. */ - close_cursor(conn, cursor_number, NULL); - } - PG_CATCH(); - { PQclear(res); - PG_RE_THROW(); + + /* Must be EOF if we didn't get all the rows requested. */ + if (numrows < fetch_size) + break; } - PG_END_TRY(); + + /* Close the cursor, just to be tidy. */ + close_cursor(conn, cursor_number, NULL); ReleaseConnection(conn); @@ -5481,7 +5381,7 @@ postgresImportForeignSchema(ImportForeignSchemaStmt *stmt, Oid serverOid) UserMapping *mapping; PGconn *conn; StringInfoData buf; - PGresult *volatile res = NULL; + PGresult *res; int numrows, i; ListCell *lc; @@ -5520,243 +5420,231 @@ postgresImportForeignSchema(ImportForeignSchemaStmt *stmt, Oid serverOid) /* Create workspace for strings */ initStringInfo(&buf); - /* In what follows, do not risk leaking any PGresults. */ - PG_TRY(); - { - /* Check that the schema really exists */ - appendStringInfoString(&buf, "SELECT 1 FROM pg_catalog.pg_namespace WHERE nspname = "); - deparseStringLiteral(&buf, stmt->remote_schema); + /* Check that the schema really exists */ + appendStringInfoString(&buf, "SELECT 1 FROM pg_catalog.pg_namespace WHERE nspname = "); + deparseStringLiteral(&buf, stmt->remote_schema); - res = pgfdw_exec_query(conn, buf.data, NULL); - if (PQresultStatus(res) != PGRES_TUPLES_OK) - pgfdw_report_error(ERROR, res, conn, false, buf.data); + res = pgfdw_exec_query(conn, buf.data, NULL); + if (PQresultStatus(res) != PGRES_TUPLES_OK) + pgfdw_report_error(res, conn, buf.data); - if (PQntuples(res) != 1) - ereport(ERROR, - (errcode(ERRCODE_FDW_SCHEMA_NOT_FOUND), - errmsg("schema \"%s\" is not present on foreign server \"%s\"", - stmt->remote_schema, server->servername))); + if (PQntuples(res) != 1) + ereport(ERROR, + (errcode(ERRCODE_FDW_SCHEMA_NOT_FOUND), + errmsg("schema \"%s\" is not present on foreign server \"%s\"", + stmt->remote_schema, server->servername))); - PQclear(res); - res = NULL; - resetStringInfo(&buf); + PQclear(res); + resetStringInfo(&buf); - /* - * Fetch all table data from this schema, possibly restricted by - * EXCEPT or LIMIT TO. (We don't actually need to pay any attention - * to EXCEPT/LIMIT TO here, because the core code will filter the - * statements we return according to those lists anyway. But it - * should save a few cycles to not process excluded tables in the - * first place.) - * - * Import table data for partitions only when they are explicitly - * specified in LIMIT TO clause. Otherwise ignore them and only - * include the definitions of the root partitioned tables to allow - * access to the complete remote data set locally in the schema - * imported. - * - * Note: because we run the connection with search_path restricted to - * pg_catalog, the format_type() and pg_get_expr() outputs will always - * include a schema name for types/functions in other schemas, which - * is what we want. - */ + /* + * Fetch all table data from this schema, possibly restricted by EXCEPT or + * LIMIT TO. (We don't actually need to pay any attention to EXCEPT/LIMIT + * TO here, because the core code will filter the statements we return + * according to those lists anyway. But it should save a few cycles to + * not process excluded tables in the first place.) + * + * Import table data for partitions only when they are explicitly + * specified in LIMIT TO clause. Otherwise ignore them and only include + * the definitions of the root partitioned tables to allow access to the + * complete remote data set locally in the schema imported. + * + * Note: because we run the connection with search_path restricted to + * pg_catalog, the format_type() and pg_get_expr() outputs will always + * include a schema name for types/functions in other schemas, which is + * what we want. + */ + appendStringInfoString(&buf, + "SELECT relname, " + " attname, " + " format_type(atttypid, atttypmod), " + " attnotnull, " + " pg_get_expr(adbin, adrelid), "); + + /* Generated columns are supported since Postgres 12 */ + if (PQserverVersion(conn) >= 120000) appendStringInfoString(&buf, - "SELECT relname, " - " attname, " - " format_type(atttypid, atttypmod), " - " attnotnull, " - " pg_get_expr(adbin, adrelid), "); - - /* Generated columns are supported since Postgres 12 */ - if (PQserverVersion(conn) >= 120000) - appendStringInfoString(&buf, - " attgenerated, "); - else - appendStringInfoString(&buf, - " NULL, "); - - if (import_collate) - appendStringInfoString(&buf, - " collname, " - " collnsp.nspname "); - else - appendStringInfoString(&buf, - " NULL, NULL "); - + " attgenerated, "); + else appendStringInfoString(&buf, - "FROM pg_class c " - " JOIN pg_namespace n ON " - " relnamespace = n.oid " - " LEFT JOIN pg_attribute a ON " - " attrelid = c.oid AND attnum > 0 " - " AND NOT attisdropped " - " LEFT JOIN pg_attrdef ad ON " - " adrelid = c.oid AND adnum = attnum "); - - if (import_collate) - appendStringInfoString(&buf, - " LEFT JOIN pg_collation coll ON " - " coll.oid = attcollation " - " LEFT JOIN pg_namespace collnsp ON " - " collnsp.oid = collnamespace "); + " NULL, "); + if (import_collate) appendStringInfoString(&buf, - "WHERE c.relkind IN (" - CppAsString2(RELKIND_RELATION) "," - CppAsString2(RELKIND_VIEW) "," - CppAsString2(RELKIND_FOREIGN_TABLE) "," - CppAsString2(RELKIND_MATVIEW) "," - CppAsString2(RELKIND_PARTITIONED_TABLE) ") " - " AND n.nspname = "); - deparseStringLiteral(&buf, stmt->remote_schema); - - /* Partitions are supported since Postgres 10 */ - if (PQserverVersion(conn) >= 100000 && - stmt->list_type != FDW_IMPORT_SCHEMA_LIMIT_TO) - appendStringInfoString(&buf, " AND NOT c.relispartition "); - - /* Apply restrictions for LIMIT TO and EXCEPT */ - if (stmt->list_type == FDW_IMPORT_SCHEMA_LIMIT_TO || - stmt->list_type == FDW_IMPORT_SCHEMA_EXCEPT) + " collname, " + " collnsp.nspname "); + else + appendStringInfoString(&buf, + " NULL, NULL "); + + appendStringInfoString(&buf, + "FROM pg_class c " + " JOIN pg_namespace n ON " + " relnamespace = n.oid " + " LEFT JOIN pg_attribute a ON " + " attrelid = c.oid AND attnum > 0 " + " AND NOT attisdropped " + " LEFT JOIN pg_attrdef ad ON " + " adrelid = c.oid AND adnum = attnum "); + + if (import_collate) + appendStringInfoString(&buf, + " LEFT JOIN pg_collation coll ON " + " coll.oid = attcollation " + " LEFT JOIN pg_namespace collnsp ON " + " collnsp.oid = collnamespace "); + + appendStringInfoString(&buf, + "WHERE c.relkind IN (" + CppAsString2(RELKIND_RELATION) "," + CppAsString2(RELKIND_VIEW) "," + CppAsString2(RELKIND_FOREIGN_TABLE) "," + CppAsString2(RELKIND_MATVIEW) "," + CppAsString2(RELKIND_PARTITIONED_TABLE) ") " + " AND n.nspname = "); + deparseStringLiteral(&buf, stmt->remote_schema); + + /* Partitions are supported since Postgres 10 */ + if (PQserverVersion(conn) >= 100000 && + stmt->list_type != FDW_IMPORT_SCHEMA_LIMIT_TO) + appendStringInfoString(&buf, " AND NOT c.relispartition "); + + /* Apply restrictions for LIMIT TO and EXCEPT */ + if (stmt->list_type == FDW_IMPORT_SCHEMA_LIMIT_TO || + stmt->list_type == FDW_IMPORT_SCHEMA_EXCEPT) + { + bool first_item = true; + + appendStringInfoString(&buf, " AND c.relname "); + if (stmt->list_type == FDW_IMPORT_SCHEMA_EXCEPT) + appendStringInfoString(&buf, "NOT "); + appendStringInfoString(&buf, "IN ("); + + /* Append list of table names within IN clause */ + foreach(lc, stmt->table_list) { - bool first_item = true; + RangeVar *rv = (RangeVar *) lfirst(lc); - appendStringInfoString(&buf, " AND c.relname "); - if (stmt->list_type == FDW_IMPORT_SCHEMA_EXCEPT) - appendStringInfoString(&buf, "NOT "); - appendStringInfoString(&buf, "IN ("); + if (first_item) + first_item = false; + else + appendStringInfoString(&buf, ", "); + deparseStringLiteral(&buf, rv->relname); + } + appendStringInfoChar(&buf, ')'); + } - /* Append list of table names within IN clause */ - foreach(lc, stmt->table_list) - { - RangeVar *rv = (RangeVar *) lfirst(lc); + /* Append ORDER BY at the end of query to ensure output ordering */ + appendStringInfoString(&buf, " ORDER BY c.relname, a.attnum"); - if (first_item) - first_item = false; - else - appendStringInfoString(&buf, ", "); - deparseStringLiteral(&buf, rv->relname); - } - appendStringInfoChar(&buf, ')'); - } + /* Fetch the data */ + res = pgfdw_exec_query(conn, buf.data, NULL); + if (PQresultStatus(res) != PGRES_TUPLES_OK) + pgfdw_report_error(res, conn, buf.data); - /* Append ORDER BY at the end of query to ensure output ordering */ - appendStringInfoString(&buf, " ORDER BY c.relname, a.attnum"); + /* Process results */ + numrows = PQntuples(res); + /* note: incrementation of i happens in inner loop's while() test */ + for (i = 0; i < numrows;) + { + char *tablename = PQgetvalue(res, i, 0); + bool first_item = true; - /* Fetch the data */ - res = pgfdw_exec_query(conn, buf.data, NULL); - if (PQresultStatus(res) != PGRES_TUPLES_OK) - pgfdw_report_error(ERROR, res, conn, false, buf.data); + resetStringInfo(&buf); + appendStringInfo(&buf, "CREATE FOREIGN TABLE %s (\n", + quote_identifier(tablename)); - /* Process results */ - numrows = PQntuples(res); - /* note: incrementation of i happens in inner loop's while() test */ - for (i = 0; i < numrows;) + /* Scan all rows for this table */ + do { - char *tablename = PQgetvalue(res, i, 0); - bool first_item = true; + char *attname; + char *typename; + char *attnotnull; + char *attgenerated; + char *attdefault; + char *collname; + char *collnamespace; + + /* If table has no columns, we'll see nulls here */ + if (PQgetisnull(res, i, 1)) + continue; - resetStringInfo(&buf); - appendStringInfo(&buf, "CREATE FOREIGN TABLE %s (\n", - quote_identifier(tablename)); + attname = PQgetvalue(res, i, 1); + typename = PQgetvalue(res, i, 2); + attnotnull = PQgetvalue(res, i, 3); + attdefault = PQgetisnull(res, i, 4) ? NULL : + PQgetvalue(res, i, 4); + attgenerated = PQgetisnull(res, i, 5) ? NULL : + PQgetvalue(res, i, 5); + collname = PQgetisnull(res, i, 6) ? NULL : + PQgetvalue(res, i, 6); + collnamespace = PQgetisnull(res, i, 7) ? NULL : + PQgetvalue(res, i, 7); + + if (first_item) + first_item = false; + else + appendStringInfoString(&buf, ",\n"); - /* Scan all rows for this table */ - do - { - char *attname; - char *typename; - char *attnotnull; - char *attgenerated; - char *attdefault; - char *collname; - char *collnamespace; - - /* If table has no columns, we'll see nulls here */ - if (PQgetisnull(res, i, 1)) - continue; + /* Print column name and type */ + appendStringInfo(&buf, " %s %s", + quote_identifier(attname), + typename); - attname = PQgetvalue(res, i, 1); - typename = PQgetvalue(res, i, 2); - attnotnull = PQgetvalue(res, i, 3); - attdefault = PQgetisnull(res, i, 4) ? NULL : - PQgetvalue(res, i, 4); - attgenerated = PQgetisnull(res, i, 5) ? NULL : - PQgetvalue(res, i, 5); - collname = PQgetisnull(res, i, 6) ? NULL : - PQgetvalue(res, i, 6); - collnamespace = PQgetisnull(res, i, 7) ? NULL : - PQgetvalue(res, i, 7); - - if (first_item) - first_item = false; - else - appendStringInfoString(&buf, ",\n"); + /* + * Add column_name option so that renaming the foreign table's + * column doesn't break the association to the underlying column. + */ + appendStringInfoString(&buf, " OPTIONS (column_name "); + deparseStringLiteral(&buf, attname); + appendStringInfoChar(&buf, ')'); - /* Print column name and type */ - appendStringInfo(&buf, " %s %s", - quote_identifier(attname), - typename); + /* Add COLLATE if needed */ + if (import_collate && collname != NULL && collnamespace != NULL) + appendStringInfo(&buf, " COLLATE %s.%s", + quote_identifier(collnamespace), + quote_identifier(collname)); - /* - * Add column_name option so that renaming the foreign table's - * column doesn't break the association to the underlying - * column. - */ - appendStringInfoString(&buf, " OPTIONS (column_name "); - deparseStringLiteral(&buf, attname); - appendStringInfoChar(&buf, ')'); - - /* Add COLLATE if needed */ - if (import_collate && collname != NULL && collnamespace != NULL) - appendStringInfo(&buf, " COLLATE %s.%s", - quote_identifier(collnamespace), - quote_identifier(collname)); - - /* Add DEFAULT if needed */ - if (import_default && attdefault != NULL && - (!attgenerated || !attgenerated[0])) - appendStringInfo(&buf, " DEFAULT %s", attdefault); - - /* Add GENERATED if needed */ - if (import_generated && attgenerated != NULL && - attgenerated[0] == ATTRIBUTE_GENERATED_STORED) - { - Assert(attdefault != NULL); - appendStringInfo(&buf, - " GENERATED ALWAYS AS (%s) STORED", - attdefault); - } + /* Add DEFAULT if needed */ + if (import_default && attdefault != NULL && + (!attgenerated || !attgenerated[0])) + appendStringInfo(&buf, " DEFAULT %s", attdefault); - /* Add NOT NULL if needed */ - if (import_not_null && attnotnull[0] == 't') - appendStringInfoString(&buf, " NOT NULL"); + /* Add GENERATED if needed */ + if (import_generated && attgenerated != NULL && + attgenerated[0] == ATTRIBUTE_GENERATED_STORED) + { + Assert(attdefault != NULL); + appendStringInfo(&buf, + " GENERATED ALWAYS AS (%s) STORED", + attdefault); } - while (++i < numrows && - strcmp(PQgetvalue(res, i, 0), tablename) == 0); - /* - * Add server name and table-level options. We specify remote - * schema and table name as options (the latter to ensure that - * renaming the foreign table doesn't break the association). - */ - appendStringInfo(&buf, "\n) SERVER %s\nOPTIONS (", - quote_identifier(server->servername)); + /* Add NOT NULL if needed */ + if (import_not_null && attnotnull[0] == 't') + appendStringInfoString(&buf, " NOT NULL"); + } + while (++i < numrows && + strcmp(PQgetvalue(res, i, 0), tablename) == 0); - appendStringInfoString(&buf, "schema_name "); - deparseStringLiteral(&buf, stmt->remote_schema); - appendStringInfoString(&buf, ", table_name "); - deparseStringLiteral(&buf, tablename); + /* + * Add server name and table-level options. We specify remote schema + * and table name as options (the latter to ensure that renaming the + * foreign table doesn't break the association). + */ + appendStringInfo(&buf, "\n) SERVER %s\nOPTIONS (", + quote_identifier(server->servername)); - appendStringInfoString(&buf, ");"); + appendStringInfoString(&buf, "schema_name "); + deparseStringLiteral(&buf, stmt->remote_schema); + appendStringInfoString(&buf, ", table_name "); + deparseStringLiteral(&buf, tablename); - commands = lappend(commands, pstrdup(buf.data)); - } - } - PG_FINALLY(); - { - PQclear(res); + appendStringInfoString(&buf, ");"); + + commands = lappend(commands, pstrdup(buf.data)); } - PG_END_TRY(); + PQclear(res); ReleaseConnection(conn); @@ -7424,7 +7312,7 @@ postgresForeignAsyncNotify(AsyncRequest *areq) /* On error, report the original query, not the FETCH. */ if (!PQconsumeInput(fsstate->conn)) - pgfdw_report_error(ERROR, NULL, fsstate->conn, false, fsstate->query); + pgfdw_report_error(NULL, fsstate->conn, fsstate->query); fetch_more_data(node); @@ -7523,7 +7411,7 @@ fetch_more_data_begin(AsyncRequest *areq) fsstate->fetch_size, fsstate->cursor_number); if (!PQsendQuery(fsstate->conn, sql)) - pgfdw_report_error(ERROR, NULL, fsstate->conn, false, fsstate->query); + pgfdw_report_error(NULL, fsstate->conn, fsstate->query); /* Remember that the request is in process */ fsstate->conn_state->pendingAreq = areq; diff --git a/contrib/postgres_fdw/postgres_fdw.h b/contrib/postgres_fdw/postgres_fdw.h index 81358f3bde7df..e69735298d78f 100644 --- a/contrib/postgres_fdw/postgres_fdw.h +++ b/contrib/postgres_fdw/postgres_fdw.h @@ -15,7 +15,7 @@ #include "foreign/foreign.h" #include "lib/stringinfo.h" -#include "libpq-fe.h" +#include "libpq/libpq-be-fe.h" #include "nodes/execnodes.h" #include "nodes/pathnodes.h" #include "utils/relcache.h" @@ -166,8 +166,10 @@ extern void do_sql_command(PGconn *conn, const char *sql); extern PGresult *pgfdw_get_result(PGconn *conn); extern PGresult *pgfdw_exec_query(PGconn *conn, const char *query, PgFdwConnState *state); -extern void pgfdw_report_error(int elevel, PGresult *res, PGconn *conn, - bool clear, const char *sql); +pg_noreturn extern void pgfdw_report_error(PGresult *res, PGconn *conn, + const char *sql); +extern void pgfdw_report(int elevel, PGresult *res, PGconn *conn, + const char *sql); /* in option.c */ extern int ExtractConnectionOptions(List *defelems, diff --git a/doc/src/sgml/func.sgml b/doc/src/sgml/func.sgml index de5b5929ee078..74a16af04ad3b 100644 --- a/doc/src/sgml/func.sgml +++ b/doc/src/sgml/func.sgml @@ -3148,8 +3148,11 @@ SELECT NOT(ROW(table.*) IS NOT NULL) FROM TABLE; -- detect at least one null in Converts the first letter of each word to upper case and the - rest to lower case. Words are sequences of alphanumeric - characters separated by non-alphanumeric characters. + rest to lower case. When using the libc locale + provider, words are sequences of alphanumeric characters separated + by non-alphanumeric characters; when using the ICU locale provider, + words are separated according to + Unicode Standard Annex #29. initcap('hi THOMAS') diff --git a/doc/src/sgml/pageinspect.sgml b/doc/src/sgml/pageinspect.sgml index 1292933366555..f5014787c783b 100644 --- a/doc/src/sgml/pageinspect.sgml +++ b/doc/src/sgml/pageinspect.sgml @@ -741,7 +741,7 @@ test=# SELECT first_tid, nbytes, tids[0:5] AS some_tids For example: test=# SELECT * FROM gist_page_opaque_info(get_raw_page('test_gist_idx', 2)); - lsn | nsn | rightlink | flags + lsn | nsn | rightlink | flags ------------+------------+-----------+-------- 0/0B5FE088 | 0/00000000 | 1 | {leaf} (1 row) diff --git a/doc/src/sgml/pgbuffercache.sgml b/doc/src/sgml/pgbuffercache.sgml index 546ace8369e28..eeb85a0e04908 100644 --- a/doc/src/sgml/pgbuffercache.sgml +++ b/doc/src/sgml/pgbuffercache.sgml @@ -19,10 +19,18 @@ pg_buffercache_pages + + pg_buffercache_numa + + pg_buffercache_summary + + pg_buffercache_usage_counts + + pg_buffercache_evict @@ -489,7 +497,7 @@ - The <structname>pg_buffercache_evict_relation</structname> Function + The <structname>pg_buffercache_evict_relation()</structname> Function The pg_buffercache_evict_relation() function is very similar to the pg_buffercache_evict() function. The @@ -507,7 +515,7 @@ - The <structname>pg_buffercache_evict_all</structname> Function + The <structname>pg_buffercache_evict_all()</structname> Function The pg_buffercache_evict_all() function is very similar to the pg_buffercache_evict() function. The diff --git a/doc/src/sgml/pgstatstatements.sgml b/doc/src/sgml/pgstatstatements.sgml index 7baa07dcdbf7f..d753de5836efb 100644 --- a/doc/src/sgml/pgstatstatements.sgml +++ b/doc/src/sgml/pgstatstatements.sgml @@ -554,6 +554,24 @@ + + + generic_plan_calls bigint + + + Number of times the statement has been executed using a generic plan + + + + + + custom_plan_calls bigint + + + Number of times the statement has been executed using a custom plan + + + stats_since timestamp with time zone diff --git a/doc/src/sgml/protocol.sgml b/doc/src/sgml/protocol.sgml index b115884acb346..e56eac8fd0fa0 100644 --- a/doc/src/sgml/protocol.sgml +++ b/doc/src/sgml/protocol.sgml @@ -537,6 +537,11 @@ The frontend should not respond to this message, but should continue listening for a ReadyForQuery message. + + The PostgreSQL server will always send this + message, but some third party backend implementations of the protocol + that don't support query cancellation are known not to. + diff --git a/doc/src/sgml/ref/create_trigger.sgml b/doc/src/sgml/ref/create_trigger.sgml index ed6d206ae7143..0d8d463479bc1 100644 --- a/doc/src/sgml/ref/create_trigger.sgml +++ b/doc/src/sgml/ref/create_trigger.sgml @@ -197,9 +197,11 @@ CREATE [ OR REPLACE ] [ CONSTRAINT ] TRIGGER name of the rows inserted, deleted, or modified by the current SQL statement. This feature lets the trigger see a global view of what the statement did, not just one row at a time. This option is only allowed for - an AFTER trigger that is not a constraint trigger; also, if - the trigger is an UPDATE trigger, it must not specify - a column_name list. + an AFTER trigger on a plain table (not a foreign table). + The trigger should not be a constraint trigger. Also, if the trigger is + an UPDATE trigger, it must not specify + a column_name list when using + this option. OLD TABLE may only be specified once, and only for a trigger that can fire on UPDATE or DELETE; it creates a transition relation containing the before-images of all rows diff --git a/doc/src/sgml/ref/pg_dump.sgml b/doc/src/sgml/ref/pg_dump.sgml index 2ae084b5fa6fc..0bc7609bdf815 100644 --- a/doc/src/sgml/ref/pg_dump.sgml +++ b/doc/src/sgml/ref/pg_dump.sgml @@ -1354,6 +1354,15 @@ PostgreSQL documentation + + + + + Dump statistics. + + + + @@ -1440,33 +1449,6 @@ PostgreSQL documentation - - - - - Dump data. This is the default. - - - - - - - - - Dump schema (data definitions). This is the default. - - - - - - - - - Dump statistics. - - - - @@ -1682,7 +1664,7 @@ CREATE DATABASE foo WITH TEMPLATE template0; - If is specified, + If is specified, pg_dump will include most optimizer statistics in the resulting dump file. However, some statistics may not be included, such as those created explicitly with or diff --git a/doc/src/sgml/ref/pg_dumpall.sgml b/doc/src/sgml/ref/pg_dumpall.sgml index 8ca68da5a5560..364442f00f28e 100644 --- a/doc/src/sgml/ref/pg_dumpall.sgml +++ b/doc/src/sgml/ref/pg_dumpall.sgml @@ -16,10 +16,7 @@ PostgreSQL documentation pg_dumpall - - - export a PostgreSQL database cluster as an SQL script or to other formats - + extract a PostgreSQL database cluster into a script file @@ -36,7 +33,7 @@ PostgreSQL documentation pg_dumpall is a utility for writing out (dumping) all PostgreSQL databases - of a cluster into an SQL script file or an archive. The output contains + of a cluster into one script file. The script file contains SQL commands that can be used as input to to restore the databases. It does this by calling for each database in the cluster. @@ -55,16 +52,11 @@ PostgreSQL documentation - Plain text SQL scripts will be written to the standard output. Use the + The SQL script will be written to the standard output. Use the / option or shell operators to redirect it into a file. - - Archives in other formats will be placed in a directory named using the - /, which is required in this case. - - pg_dumpall needs to connect several times to the PostgreSQL server (once per @@ -129,85 +121,10 @@ PostgreSQL documentation Send output to the specified file. If this is omitted, the standard output is used. - Note: This option can only be omitted when is plain - - - - - - Specify the format of dump files. In plain format, all the dump data is - sent in a single text stream. This is the default. - - In all other modes, pg_dumpall first creates two files: - global.dat and map.dat, in the directory - specified by . - The first file contains global data, such as roles and tablespaces. The second - contains a mapping between database oids and names. These files are used by - pg_restore. Data for individual databases is placed in - databases subdirectory, named using the database's oid. - - - - d - directory - - - Output directory-format archives for each database, - suitable for input into pg_restore. The directory - will have database oid as its name. - - - - - - p - plain - - - Output a plain-text SQL script file (the default). - - - - - - c - custom - - - Output a custom-format archive for each database, - suitable for input into pg_restore. The archive - will be named dboid.dmp where dboid is the - oid of the database. - - - - - - t - tar - - - Output a tar-format archive for each database, - suitable for input into pg_restore. The archive - will be named dboid.tar where dboid is the - oid of the database. - - - - - - - Note: see for details - of how the various non plain text archives work. - - - - - @@ -688,6 +605,15 @@ exclude database PATTERN + + + + + Dump statistics. + + + + @@ -723,33 +649,6 @@ exclude database PATTERN - - - - - Dump data. This is the default. - - - - - - - - - Dump schema (data definitions). This is the default. - - - - - - - - - Dump statistics. - - - - @@ -961,7 +860,7 @@ exclude database PATTERN - If is specified, + If is specified, pg_dumpall will include most optimizer statistics in the resulting dump file. However, some statistics may not be included, such as those created explicitly with or diff --git a/doc/src/sgml/ref/pg_restore.sgml b/doc/src/sgml/ref/pg_restore.sgml index b649bd3a5ae0f..261ead1503955 100644 --- a/doc/src/sgml/ref/pg_restore.sgml +++ b/doc/src/sgml/ref/pg_restore.sgml @@ -18,9 +18,8 @@ PostgreSQL documentation pg_restore - restore PostgreSQL databases from archives - created by pg_dump or - pg_dumpall + restore a PostgreSQL database from an + archive file created by pg_dump @@ -39,14 +38,13 @@ PostgreSQL documentation pg_restore is a utility for restoring a - PostgreSQL database or cluster from an archive - created by or - in one of the non-plain-text + PostgreSQL database from an archive + created by in one of the non-plain-text formats. It will issue the commands necessary to reconstruct the - database or cluster to the state it was in at the time it was saved. The - archives also allow pg_restore to + database to the state it was in at the time it was saved. The + archive files also allow pg_restore to be selective about what is restored, or even to reorder the items - prior to being restored. The archive formats are designed to be + prior to being restored. The archive files are designed to be portable across architectures. @@ -54,17 +52,10 @@ PostgreSQL documentation pg_restore can operate in two modes. If a database name is specified, pg_restore connects to that database and restores archive contents directly into - the database. - When restoring from a dump made by pg_dumpall, - each database will be created and then the restoration will be run in that - database. - - Otherwise, when a database name is not specified, a script containing the SQL - commands necessary to rebuild the database or cluster is created and written + the database. Otherwise, a script containing the SQL + commands necessary to rebuild the database is created and written to a file or standard output. This script output is equivalent to - the plain text output format of pg_dump or - pg_dumpall. - + the plain text output format of pg_dump. Some of the options controlling the output are therefore analogous to pg_dump options. @@ -149,8 +140,6 @@ PostgreSQL documentation commands that mention this database. Access privileges for the database itself are also restored, unless is specified. - is required when restoring multiple databases - from an archive created by pg_dumpall. @@ -246,19 +235,6 @@ PostgreSQL documentation - - - - - - Restore only global objects (roles and tablespaces), no databases. - - - This option is only relevant when restoring from an archive made using pg_dumpall. - - - - @@ -603,28 +579,6 @@ PostgreSQL documentation - - - - - Do not restore databases whose name matches - pattern. - Multiple patterns can be excluded by writing multiple - switches. The - pattern parameter is - interpreted as a pattern according to the same rules used by - psql's \d - commands (see ), - so multiple databases can also be excluded by writing wildcard - characters in the pattern. When using wildcards, be careful to - quote the pattern if needed to prevent shell wildcard expansion. - - - This option is only relevant when restoring from an archive made using pg_dumpall. - - - - @@ -861,6 +815,16 @@ PostgreSQL documentation + + + + + Output commands to restore statistics, if the archive contains them. + This is the default. + + + + @@ -919,36 +883,6 @@ PostgreSQL documentation - - - - - Output commands to restore data, if the archive contains them. - This is the default. - - - - - - - - - Output commands to restore schema (data definitions), if the archive - contains them. This is the default. - - - - - - - - - Output commands to restore statistics, if the archive contains them. - This is the default. - - - - diff --git a/doc/src/sgml/ref/vacuumdb.sgml b/doc/src/sgml/ref/vacuumdb.sgml index b0680a61814cc..c7d9dca17b867 100644 --- a/doc/src/sgml/ref/vacuumdb.sgml +++ b/doc/src/sgml/ref/vacuumdb.sgml @@ -282,9 +282,11 @@ PostgreSQL documentation Only analyze relations that are missing statistics for a column, index - expression, or extended statistics object. This option prevents - vacuumdb from deleting existing statistics - so that the query optimizer's choices do not become transiently worse. + expression, or extended statistics object. When used with + , this option prevents + vacuumdb from temporarily replacing existing + statistics with ones generated with lower statistics targets, thus + avoiding transiently worse query optimizer choices. This option can only be used in conjunction with diff --git a/meson.build b/meson.build index 5365aaf95e64b..ca423dc8e12f3 100644 --- a/meson.build +++ b/meson.build @@ -1996,10 +1996,7 @@ if cc.links(''' cdata.set('HAVE__BUILTIN_OP_OVERFLOW', 1) endif - -# XXX: The configure.ac check for __cpuid() is broken, we don't copy that -# here. To prevent problems due to two detection methods working, stop -# checking after one. +# Check for __get_cpuid() and __cpuid(). if cc.links(''' #include int main(int arg, char **argv) diff --git a/src/Makefile.global.in b/src/Makefile.global.in index 04952b533ded9..8b1b357beaa04 100644 --- a/src/Makefile.global.in +++ b/src/Makefile.global.in @@ -254,7 +254,7 @@ CPP = @CPP@ CPPFLAGS = @CPPFLAGS@ PG_SYSROOT = @PG_SYSROOT@ -override CPPFLAGS := $(ICU_CFLAGS) $(LIBNUMA_CFLAGS) $(LIBURING_CFLAGS) $(CPPFLAGS) +override CPPFLAGS += $(ICU_CFLAGS) $(LIBNUMA_CFLAGS) $(LIBURING_CFLAGS) ifdef PGXS override CPPFLAGS := -I$(includedir_server) -I$(includedir_internal) $(CPPFLAGS) diff --git a/src/backend/access/heap/visibilitymap.c b/src/backend/access/heap/visibilitymap.c index 745a04ef26e29..8f918e00af7ed 100644 --- a/src/backend/access/heap/visibilitymap.c +++ b/src/backend/access/heap/visibilitymap.c @@ -364,7 +364,7 @@ visibilitymap_get_status(Relation rel, BlockNumber heapBlk, Buffer *vmbuf) { *vmbuf = vm_readbuf(rel, mapBlock, false); if (!BufferIsValid(*vmbuf)) - return false; + return (uint8) 0; } map = PageGetContents(BufferGetPage(*vmbuf)); diff --git a/src/backend/access/index/amapi.c b/src/backend/access/index/amapi.c index f0f4f974bcedb..60684c5342279 100644 --- a/src/backend/access/index/amapi.c +++ b/src/backend/access/index/amapi.c @@ -42,6 +42,19 @@ GetIndexAmRoutine(Oid amhandler) elog(ERROR, "index access method handler function %u did not return an IndexAmRoutine struct", amhandler); + /* Assert that all required callbacks are present. */ + Assert(routine->ambuild != NULL); + Assert(routine->ambuildempty != NULL); + Assert(routine->aminsert != NULL); + Assert(routine->ambulkdelete != NULL); + Assert(routine->amvacuumcleanup != NULL); + Assert(routine->amcostestimate != NULL); + Assert(routine->amoptions != NULL); + Assert(routine->amvalidate != NULL); + Assert(routine->ambeginscan != NULL); + Assert(routine->amrescan != NULL); + Assert(routine->amendscan != NULL); + return routine; } diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c index eefffc4277a1a..5553c20fee8ce 100644 --- a/src/backend/access/transam/xlog.c +++ b/src/backend/access/transam/xlog.c @@ -96,6 +96,7 @@ #include "utils/guc_hooks.h" #include "utils/guc_tables.h" #include "utils/injection_point.h" +#include "utils/pgstat_internal.h" #include "utils/ps_status.h" #include "utils/relmapper.h" #include "utils/snapmgr.h" @@ -702,7 +703,7 @@ static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version); static void WriteControlFile(void); static void ReadControlFile(void); static void UpdateControlFile(void); -static char *str_time(pg_time_t tnow); +static char *str_time(pg_time_t tnow, char *buf, size_t bufsize); static int get_sync_bit(int method); @@ -1091,6 +1092,9 @@ XLogInsertRecord(XLogRecData *rdata, pgWalUsage.wal_bytes += rechdr->xl_tot_len; pgWalUsage.wal_records++; pgWalUsage.wal_fpi += num_fpi; + + /* Required for the flush of pending stats WAL data */ + pgstat_report_fixed = true; } return EndPos; @@ -2108,6 +2112,12 @@ AdvanceXLInsertBuffer(XLogRecPtr upto, TimeLineID tli, bool opportunistic) LWLockRelease(WALWriteLock); pgWalUsage.wal_buffers_full++; TRACE_POSTGRESQL_WAL_BUFFER_WRITE_DIRTY_DONE(); + + /* + * Required for the flush of pending stats WAL data, per + * update of pgWalUsage. + */ + pgstat_report_fixed = true; } } } @@ -5361,11 +5371,9 @@ BootStrapXLOG(uint32 data_checksum_version) } static char * -str_time(pg_time_t tnow) +str_time(pg_time_t tnow, char *buf, size_t bufsize) { - char *buf = palloc(128); - - pg_strftime(buf, 128, + pg_strftime(buf, bufsize, "%Y-%m-%d %H:%M:%S %Z", pg_localtime(&tnow, log_timezone)); @@ -5608,6 +5616,7 @@ StartupXLOG(void) XLogRecPtr missingContrecPtr; TransactionId oldestActiveXID; bool promoted = false; + char timebuf[128]; /* * We should have an aux process resource owner to use, and we should not @@ -5636,25 +5645,29 @@ StartupXLOG(void) */ ereport(IsPostmasterEnvironment ? LOG : NOTICE, (errmsg("database system was shut down at %s", - str_time(ControlFile->time)))); + str_time(ControlFile->time, + timebuf, sizeof(timebuf))))); break; case DB_SHUTDOWNED_IN_RECOVERY: ereport(LOG, (errmsg("database system was shut down in recovery at %s", - str_time(ControlFile->time)))); + str_time(ControlFile->time, + timebuf, sizeof(timebuf))))); break; case DB_SHUTDOWNING: ereport(LOG, (errmsg("database system shutdown was interrupted; last known up at %s", - str_time(ControlFile->time)))); + str_time(ControlFile->time, + timebuf, sizeof(timebuf))))); break; case DB_IN_CRASH_RECOVERY: ereport(LOG, (errmsg("database system was interrupted while in recovery at %s", - str_time(ControlFile->time)), + str_time(ControlFile->time, + timebuf, sizeof(timebuf))), errhint("This probably means that some data is corrupted and" " you will have to use the last backup for recovery."))); break; @@ -5662,7 +5675,8 @@ StartupXLOG(void) case DB_IN_ARCHIVE_RECOVERY: ereport(LOG, (errmsg("database system was interrupted while in recovery at log time %s", - str_time(ControlFile->checkPointCopy.time)), + str_time(ControlFile->checkPointCopy.time, + timebuf, sizeof(timebuf))), errhint("If this has occurred more than once some data might be corrupted" " and you might need to choose an earlier recovery target."))); break; @@ -5670,7 +5684,8 @@ StartupXLOG(void) case DB_IN_PRODUCTION: ereport(LOG, (errmsg("database system was interrupted; last known up at %s", - str_time(ControlFile->time)))); + str_time(ControlFile->time, + timebuf, sizeof(timebuf))))); break; default: @@ -6315,6 +6330,12 @@ StartupXLOG(void) */ CompleteCommitTsInitialization(); + /* Clean up EndOfWalRecoveryInfo data to appease Valgrind leak checking */ + if (endOfRecoveryInfo->lastPage) + pfree(endOfRecoveryInfo->lastPage); + pfree(endOfRecoveryInfo->recoveryStopReason); + pfree(endOfRecoveryInfo); + /* * All done with end-of-recovery actions. * diff --git a/src/backend/access/transam/xlogrecovery.c b/src/backend/access/transam/xlogrecovery.c index e8f3ba00caae7..f23ec8969c27d 100644 --- a/src/backend/access/transam/xlogrecovery.c +++ b/src/backend/access/transam/xlogrecovery.c @@ -1626,6 +1626,7 @@ ShutdownWalRecovery(void) close(readFile); readFile = -1; } + pfree(xlogreader->private_data); XLogReaderFree(xlogreader); XLogPrefetcherFree(xlogprefetcher); diff --git a/src/backend/catalog/pg_subscription.c b/src/backend/catalog/pg_subscription.c index 63c2992d19f75..244acf52f3602 100644 --- a/src/backend/catalog/pg_subscription.c +++ b/src/backend/catalog/pg_subscription.c @@ -320,7 +320,7 @@ AddSubscriptionRelState(Oid subid, Oid relid, char state, */ void UpdateSubscriptionRelState(Oid subid, Oid relid, char state, - XLogRecPtr sublsn) + XLogRecPtr sublsn, bool already_locked) { Relation rel; HeapTuple tup; @@ -328,9 +328,24 @@ UpdateSubscriptionRelState(Oid subid, Oid relid, char state, Datum values[Natts_pg_subscription_rel]; bool replaces[Natts_pg_subscription_rel]; - LockSharedObject(SubscriptionRelationId, subid, 0, AccessShareLock); + if (already_locked) + { +#ifdef USE_ASSERT_CHECKING + LOCKTAG tag; - rel = table_open(SubscriptionRelRelationId, RowExclusiveLock); + Assert(CheckRelationOidLockedByMe(SubscriptionRelRelationId, + RowExclusiveLock, true)); + SET_LOCKTAG_OBJECT(tag, InvalidOid, SubscriptionRelationId, subid, 0); + Assert(LockHeldByMe(&tag, AccessShareLock, true)); +#endif + + rel = table_open(SubscriptionRelRelationId, NoLock); + } + else + { + LockSharedObject(SubscriptionRelationId, subid, 0, AccessShareLock); + rel = table_open(SubscriptionRelRelationId, RowExclusiveLock); + } /* Try finding existing mapping. */ tup = SearchSysCacheCopy2(SUBSCRIPTIONRELMAP, diff --git a/src/backend/commands/explain.c b/src/backend/commands/explain.c index 7e2792ead715b..8345bc0264b23 100644 --- a/src/backend/commands/explain.c +++ b/src/backend/commands/explain.c @@ -3582,6 +3582,7 @@ static void show_memoize_info(MemoizeState *mstate, List *ancestors, ExplainState *es) { Plan *plan = ((PlanState *) mstate)->plan; + Memoize *mplan = (Memoize *) plan; ListCell *lc; List *context; StringInfoData keystr; @@ -3602,7 +3603,7 @@ show_memoize_info(MemoizeState *mstate, List *ancestors, ExplainState *es) plan, ancestors); - foreach(lc, ((Memoize *) plan)->param_exprs) + foreach(lc, mplan->param_exprs) { Node *expr = (Node *) lfirst(lc); @@ -3618,6 +3619,24 @@ show_memoize_info(MemoizeState *mstate, List *ancestors, ExplainState *es) pfree(keystr.data); + if (es->costs) + { + if (es->format == EXPLAIN_FORMAT_TEXT) + { + ExplainIndentText(es); + appendStringInfo(es->str, "Estimates: capacity=%u distinct keys=%.0f lookups=%.0f hit percent=%.2f%%\n", + mplan->est_entries, mplan->est_unique_keys, + mplan->est_calls, mplan->est_hit_ratio * 100.0); + } + else + { + ExplainPropertyUInteger("Estimated Capacity", NULL, mplan->est_entries, es); + ExplainPropertyFloat("Estimated Distinct Lookup Keys", NULL, mplan->est_unique_keys, 0, es); + ExplainPropertyFloat("Estimated Lookups", NULL, mplan->est_calls, 0, es); + ExplainPropertyFloat("Estimated Hit Percent", NULL, mplan->est_hit_ratio * 100.0, 2, es); + } + } + if (!es->analyze) return; diff --git a/src/backend/commands/foreigncmds.c b/src/backend/commands/foreigncmds.c index 8d2d743154462..77f8461f42eee 100644 --- a/src/backend/commands/foreigncmds.c +++ b/src/backend/commands/foreigncmds.c @@ -1588,6 +1588,7 @@ ImportForeignSchema(ImportForeignSchemaStmt *stmt) pstmt->utilityStmt = (Node *) cstmt; pstmt->stmt_location = rs->stmt_location; pstmt->stmt_len = rs->stmt_len; + pstmt->planOrigin = PLAN_STMT_INTERNAL; /* Execute statement */ ProcessUtility(pstmt, cmd, false, diff --git a/src/backend/commands/schemacmds.c b/src/backend/commands/schemacmds.c index 546160f09410e..0f03d9743d203 100644 --- a/src/backend/commands/schemacmds.c +++ b/src/backend/commands/schemacmds.c @@ -215,6 +215,7 @@ CreateSchemaCommand(CreateSchemaStmt *stmt, const char *queryString, wrapper->utilityStmt = stmt; wrapper->stmt_location = stmt_location; wrapper->stmt_len = stmt_len; + wrapper->planOrigin = PLAN_STMT_INTERNAL; /* do this step */ ProcessUtility(wrapper, diff --git a/src/backend/executor/execParallel.c b/src/backend/executor/execParallel.c index f3e77bda27906..f098a5557cf07 100644 --- a/src/backend/executor/execParallel.c +++ b/src/backend/executor/execParallel.c @@ -189,6 +189,7 @@ ExecSerializePlan(Plan *plan, EState *estate) pstmt->permInfos = estate->es_rteperminfos; pstmt->resultRelations = NIL; pstmt->appendRelations = NIL; + pstmt->planOrigin = PLAN_STMT_INTERNAL; /* * Transfer only parallel-safe subplans, leaving a NULL "hole" in the list diff --git a/src/backend/jit/llvm/Makefile b/src/backend/jit/llvm/Makefile index e8c12060b93df..68677ba42e189 100644 --- a/src/backend/jit/llvm/Makefile +++ b/src/backend/jit/llvm/Makefile @@ -31,7 +31,7 @@ endif # All files in this directory use LLVM. CFLAGS += $(LLVM_CFLAGS) CXXFLAGS += $(LLVM_CXXFLAGS) -override CPPFLAGS := $(LLVM_CPPFLAGS) $(CPPFLAGS) +override CPPFLAGS += $(LLVM_CPPFLAGS) SHLIB_LINK += $(LLVM_LIBS) # Because this module includes C++ files, we need to use a C++ diff --git a/src/backend/libpq/auth.c b/src/backend/libpq/auth.c index 9f4d05ffbd453..4da46666439db 100644 --- a/src/backend/libpq/auth.c +++ b/src/backend/libpq/auth.c @@ -94,8 +94,16 @@ static int auth_peer(hbaPort *port); #define PGSQL_PAM_SERVICE "postgresql" /* Service name passed to PAM */ +/* Work around original Solaris' lack of "const" in the conv_proc signature */ +#ifdef _PAM_LEGACY_NONCONST +#define PG_PAM_CONST +#else +#define PG_PAM_CONST const +#endif + static int CheckPAMAuth(Port *port, const char *user, const char *password); -static int pam_passwd_conv_proc(int num_msg, const struct pam_message **msg, +static int pam_passwd_conv_proc(int num_msg, + PG_PAM_CONST struct pam_message **msg, struct pam_response **resp, void *appdata_ptr); static struct pam_conv pam_passw_conv = { @@ -1917,7 +1925,7 @@ auth_peer(hbaPort *port) */ static int -pam_passwd_conv_proc(int num_msg, const struct pam_message **msg, +pam_passwd_conv_proc(int num_msg, PG_PAM_CONST struct pam_message **msg, struct pam_response **resp, void *appdata_ptr) { const char *passwd; diff --git a/src/backend/libpq/pqcomm.c b/src/backend/libpq/pqcomm.c index e5171467de18d..25f739a6a17d4 100644 --- a/src/backend/libpq/pqcomm.c +++ b/src/backend/libpq/pqcomm.c @@ -858,7 +858,6 @@ RemoveSocketFiles(void) (void) unlink(sock_path); } /* Since we're about to exit, no need to reclaim storage */ - sock_paths = NIL; } diff --git a/src/backend/libpq/pqmq.c b/src/backend/libpq/pqmq.c index f1a08bc32ca17..5f39949a36773 100644 --- a/src/backend/libpq/pqmq.c +++ b/src/backend/libpq/pqmq.c @@ -23,7 +23,7 @@ #include "tcop/tcopprot.h" #include "utils/builtins.h" -static shm_mq_handle *pq_mq_handle; +static shm_mq_handle *pq_mq_handle = NULL; static bool pq_mq_busy = false; static pid_t pq_mq_parallel_leader_pid = 0; static ProcNumber pq_mq_parallel_leader_proc_number = INVALID_PROC_NUMBER; @@ -66,7 +66,11 @@ pq_redirect_to_shm_mq(dsm_segment *seg, shm_mq_handle *mqh) static void pq_cleanup_redirect_to_shm_mq(dsm_segment *seg, Datum arg) { - pq_mq_handle = NULL; + if (pq_mq_handle != NULL) + { + pfree(pq_mq_handle); + pq_mq_handle = NULL; + } whereToSendOutput = DestNone; } @@ -131,8 +135,11 @@ mq_putmessage(char msgtype, const char *s, size_t len) if (pq_mq_busy) { if (pq_mq_handle != NULL) + { shm_mq_detach(pq_mq_handle); - pq_mq_handle = NULL; + pfree(pq_mq_handle); + pq_mq_handle = NULL; + } return EOF; } @@ -152,8 +159,6 @@ mq_putmessage(char msgtype, const char *s, size_t len) iov[1].data = s; iov[1].len = len; - Assert(pq_mq_handle != NULL); - for (;;) { /* @@ -161,6 +166,7 @@ mq_putmessage(char msgtype, const char *s, size_t len) * that the shared memory value is updated before we send the parallel * message signal right after this. */ + Assert(pq_mq_handle != NULL); result = shm_mq_sendv(pq_mq_handle, iov, 2, true, true); if (pq_mq_parallel_leader_pid != 0) diff --git a/src/backend/optimizer/path/costsize.c b/src/backend/optimizer/path/costsize.c index 1f04a2c182ca9..344a3188317b1 100644 --- a/src/backend/optimizer/path/costsize.c +++ b/src/backend/optimizer/path/costsize.c @@ -2572,13 +2572,13 @@ cost_memoize_rescan(PlannerInfo *root, MemoizePath *mpath, Cost input_startup_cost = mpath->subpath->startup_cost; Cost input_total_cost = mpath->subpath->total_cost; double tuples = mpath->subpath->rows; - double calls = mpath->calls; + Cardinality est_calls = mpath->est_calls; int width = mpath->subpath->pathtarget->width; double hash_mem_bytes; double est_entry_bytes; - double est_cache_entries; - double ndistinct; + Cardinality est_cache_entries; + Cardinality ndistinct; double evict_ratio; double hit_ratio; Cost startup_cost; @@ -2604,7 +2604,7 @@ cost_memoize_rescan(PlannerInfo *root, MemoizePath *mpath, est_cache_entries = floor(hash_mem_bytes / est_entry_bytes); /* estimate on the distinct number of parameter values */ - ndistinct = estimate_num_groups(root, mpath->param_exprs, calls, NULL, + ndistinct = estimate_num_groups(root, mpath->param_exprs, est_calls, NULL, &estinfo); /* @@ -2616,7 +2616,10 @@ cost_memoize_rescan(PlannerInfo *root, MemoizePath *mpath, * certainly mean a MemoizePath will never survive add_path(). */ if ((estinfo.flags & SELFLAG_USED_DEFAULT) != 0) - ndistinct = calls; + ndistinct = est_calls; + + /* Remember the ndistinct estimate for EXPLAIN */ + mpath->est_unique_keys = ndistinct; /* * Since we've already estimated the maximum number of entries we can @@ -2644,9 +2647,12 @@ cost_memoize_rescan(PlannerInfo *root, MemoizePath *mpath, * must look at how many scans are estimated in total for this node and * how many of those scans we expect to get a cache hit. */ - hit_ratio = ((calls - ndistinct) / calls) * + hit_ratio = ((est_calls - ndistinct) / est_calls) * (est_cache_entries / Max(ndistinct, est_cache_entries)); + /* Remember the hit ratio estimate for EXPLAIN */ + mpath->est_hit_ratio = hit_ratio; + Assert(hit_ratio >= 0 && hit_ratio <= 1.0); /* diff --git a/src/backend/optimizer/plan/createplan.c b/src/backend/optimizer/plan/createplan.c index 8a9f1d7a943a8..bfefc7dbea106 100644 --- a/src/backend/optimizer/plan/createplan.c +++ b/src/backend/optimizer/plan/createplan.c @@ -284,7 +284,10 @@ static Material *make_material(Plan *lefttree); static Memoize *make_memoize(Plan *lefttree, Oid *hashoperators, Oid *collations, List *param_exprs, bool singlerow, bool binary_mode, - uint32 est_entries, Bitmapset *keyparamids); + uint32 est_entries, Bitmapset *keyparamids, + Cardinality est_calls, + Cardinality est_unique_keys, + double est_hit_ratio); static WindowAgg *make_windowagg(List *tlist, WindowClause *wc, int partNumCols, AttrNumber *partColIdx, Oid *partOperators, Oid *partCollations, int ordNumCols, AttrNumber *ordColIdx, Oid *ordOperators, Oid *ordCollations, @@ -1753,7 +1756,8 @@ create_memoize_plan(PlannerInfo *root, MemoizePath *best_path, int flags) plan = make_memoize(subplan, operators, collations, param_exprs, best_path->singlerow, best_path->binary_mode, - best_path->est_entries, keyparamids); + best_path->est_entries, keyparamids, best_path->est_calls, + best_path->est_unique_keys, best_path->est_hit_ratio); copy_generic_path_info(&plan->plan, (Path *) best_path); @@ -6749,7 +6753,9 @@ materialize_finished_plan(Plan *subplan) static Memoize * make_memoize(Plan *lefttree, Oid *hashoperators, Oid *collations, List *param_exprs, bool singlerow, bool binary_mode, - uint32 est_entries, Bitmapset *keyparamids) + uint32 est_entries, Bitmapset *keyparamids, + Cardinality est_calls, Cardinality est_unique_keys, + double est_hit_ratio) { Memoize *node = makeNode(Memoize); Plan *plan = &node->plan; @@ -6767,6 +6773,9 @@ make_memoize(Plan *lefttree, Oid *hashoperators, Oid *collations, node->binary_mode = binary_mode; node->est_entries = est_entries; node->keyparamids = keyparamids; + node->est_calls = est_calls; + node->est_unique_keys = est_unique_keys; + node->est_hit_ratio = est_hit_ratio; return node; } diff --git a/src/backend/optimizer/plan/planner.c b/src/backend/optimizer/plan/planner.c index c989e72cac5cf..d59d6e4c6a021 100644 --- a/src/backend/optimizer/plan/planner.c +++ b/src/backend/optimizer/plan/planner.c @@ -558,6 +558,7 @@ standard_planner(Query *parse, const char *query_string, int cursorOptions, result->commandType = parse->commandType; result->queryId = parse->queryId; + result->planOrigin = PLAN_STMT_STANDARD; result->hasReturning = (parse->returningList != NIL); result->hasModifyingCTE = parse->hasModifyingCTE; result->canSetTag = parse->canSetTag; diff --git a/src/backend/optimizer/util/pathnode.c b/src/backend/optimizer/util/pathnode.c index 9cc602788eaae..a4c5867cdcb84 100644 --- a/src/backend/optimizer/util/pathnode.c +++ b/src/backend/optimizer/util/pathnode.c @@ -1689,7 +1689,7 @@ create_material_path(RelOptInfo *rel, Path *subpath) MemoizePath * create_memoize_path(PlannerInfo *root, RelOptInfo *rel, Path *subpath, List *param_exprs, List *hash_operators, - bool singlerow, bool binary_mode, double calls) + bool singlerow, bool binary_mode, Cardinality est_calls) { MemoizePath *pathnode = makeNode(MemoizePath); @@ -1710,7 +1710,6 @@ create_memoize_path(PlannerInfo *root, RelOptInfo *rel, Path *subpath, pathnode->param_exprs = param_exprs; pathnode->singlerow = singlerow; pathnode->binary_mode = binary_mode; - pathnode->calls = clamp_row_est(calls); /* * For now we set est_entries to 0. cost_memoize_rescan() does all the @@ -1720,6 +1719,12 @@ create_memoize_path(PlannerInfo *root, RelOptInfo *rel, Path *subpath, */ pathnode->est_entries = 0; + pathnode->est_calls = clamp_row_est(est_calls); + + /* These will also be set later in cost_memoize_rescan() */ + pathnode->est_unique_keys = 0.0; + pathnode->est_hit_ratio = 0.0; + /* we should not generate this path type when enable_memoize=false */ Assert(enable_memoize); pathnode->path.disabled_nodes = subpath->disabled_nodes; @@ -4259,7 +4264,7 @@ reparameterize_path(PlannerInfo *root, Path *path, mpath->hash_operators, mpath->singlerow, mpath->binary_mode, - mpath->calls); + mpath->est_calls); } default: break; diff --git a/src/backend/parser/gram.y b/src/backend/parser/gram.y index 73345bb3c7045..db43034b9db57 100644 --- a/src/backend/parser/gram.y +++ b/src/backend/parser/gram.y @@ -318,6 +318,11 @@ static Node *makeRecursiveViewSelect(char *relname, List *aliases, Node *query); %type opt_qualified_name %type opt_concurrently %type opt_drop_behavior +%type opt_utility_option_list +%type utility_option_list +%type utility_option_elem +%type utility_option_name +%type utility_option_arg %type alter_column_default opclass_item opclass_drop alter_using %type add_drop opt_asc_desc opt_nulls_order @@ -338,10 +343,6 @@ static Node *makeRecursiveViewSelect(char *relname, List *aliases, Node *query); create_extension_opt_item alter_extension_opt_item %type opt_lock lock_type cast_context -%type utility_option_name -%type utility_option_elem -%type utility_option_list -%type utility_option_arg %type drop_option %type opt_or_replace opt_no opt_grant_grant_option @@ -556,7 +557,6 @@ static Node *makeRecursiveViewSelect(char *relname, List *aliases, Node *query); %type generic_option_list alter_generic_option_list %type reindex_target_relation reindex_target_all -%type opt_reindex_option_list %type copy_generic_opt_arg copy_generic_opt_arg_list_item %type copy_generic_opt_elem @@ -1141,6 +1141,41 @@ opt_drop_behavior: | /* EMPTY */ { $$ = DROP_RESTRICT; /* default */ } ; +opt_utility_option_list: + '(' utility_option_list ')' { $$ = $2; } + | /* EMPTY */ { $$ = NULL; } + ; + +utility_option_list: + utility_option_elem + { + $$ = list_make1($1); + } + | utility_option_list ',' utility_option_elem + { + $$ = lappend($1, $3); + } + ; + +utility_option_elem: + utility_option_name utility_option_arg + { + $$ = makeDefElem($1, $2, @1); + } + ; + +utility_option_name: + NonReservedWord { $$ = $1; } + | analyze_keyword { $$ = "analyze"; } + | FORMAT_LA { $$ = "format"; } + ; + +utility_option_arg: + opt_boolean_or_string { $$ = (Node *) makeString($1); } + | NumericOnly { $$ = (Node *) $1; } + | /* EMPTY */ { $$ = NULL; } + ; + /***************************************************************************** * * CALL statement @@ -2028,18 +2063,12 @@ constraints_set_mode: * Checkpoint statement */ CheckPointStmt: - CHECKPOINT + CHECKPOINT opt_utility_option_list { CheckPointStmt *n = makeNode(CheckPointStmt); $$ = (Node *) n; - } - | CHECKPOINT '(' utility_option_list ')' - { - CheckPointStmt *n = makeNode(CheckPointStmt); - - $$ = (Node *) n; - n->options = $3; + n->options = $2; } ; @@ -9354,7 +9383,7 @@ DropTransformStmt: DROP TRANSFORM opt_if_exists FOR Typename LANGUAGE name opt_d *****************************************************************************/ ReindexStmt: - REINDEX opt_reindex_option_list reindex_target_relation opt_concurrently qualified_name + REINDEX opt_utility_option_list reindex_target_relation opt_concurrently qualified_name { ReindexStmt *n = makeNode(ReindexStmt); @@ -9367,7 +9396,7 @@ ReindexStmt: makeDefElem("concurrently", NULL, @4)); $$ = (Node *) n; } - | REINDEX opt_reindex_option_list SCHEMA opt_concurrently name + | REINDEX opt_utility_option_list SCHEMA opt_concurrently name { ReindexStmt *n = makeNode(ReindexStmt); @@ -9380,7 +9409,7 @@ ReindexStmt: makeDefElem("concurrently", NULL, @4)); $$ = (Node *) n; } - | REINDEX opt_reindex_option_list reindex_target_all opt_concurrently opt_single_name + | REINDEX opt_utility_option_list reindex_target_all opt_concurrently opt_single_name { ReindexStmt *n = makeNode(ReindexStmt); @@ -9402,10 +9431,6 @@ reindex_target_all: SYSTEM_P { $$ = REINDEX_OBJECT_SYSTEM; } | DATABASE { $$ = REINDEX_OBJECT_DATABASE; } ; -opt_reindex_option_list: - '(' utility_option_list ')' { $$ = $2; } - | /* EMPTY */ { $$ = NULL; } - ; /***************************************************************************** * @@ -11903,13 +11928,13 @@ ClusterStmt: n->params = $3; $$ = (Node *) n; } - | CLUSTER '(' utility_option_list ')' + | CLUSTER opt_utility_option_list { ClusterStmt *n = makeNode(ClusterStmt); n->relation = NULL; n->indexname = NULL; - n->params = $3; + n->params = $2; $$ = (Node *) n; } /* unparenthesized VERBOSE kept for pre-14 compatibility */ @@ -11919,21 +11944,18 @@ ClusterStmt: n->relation = $3; n->indexname = $4; - n->params = NIL; if ($2) - n->params = lappend(n->params, makeDefElem("verbose", NULL, @2)); + n->params = list_make1(makeDefElem("verbose", NULL, @2)); $$ = (Node *) n; } /* unparenthesized VERBOSE kept for pre-17 compatibility */ - | CLUSTER opt_verbose + | CLUSTER VERBOSE { ClusterStmt *n = makeNode(ClusterStmt); n->relation = NULL; n->indexname = NULL; - n->params = NIL; - if ($2) - n->params = lappend(n->params, makeDefElem("verbose", NULL, @2)); + n->params = list_make1(makeDefElem("verbose", NULL, @2)); $$ = (Node *) n; } /* kept for pre-8.3 compatibility */ @@ -11943,9 +11965,8 @@ ClusterStmt: n->relation = $5; n->indexname = $3; - n->params = NIL; if ($2) - n->params = lappend(n->params, makeDefElem("verbose", NULL, @2)); + n->params = list_make1(makeDefElem("verbose", NULL, @2)); $$ = (Node *) n; } ; @@ -11996,64 +12017,31 @@ VacuumStmt: VACUUM opt_full opt_freeze opt_verbose opt_analyze opt_vacuum_relati } ; -AnalyzeStmt: analyze_keyword opt_verbose opt_vacuum_relation_list +AnalyzeStmt: analyze_keyword opt_utility_option_list opt_vacuum_relation_list { VacuumStmt *n = makeNode(VacuumStmt); - n->options = NIL; - if ($2) - n->options = lappend(n->options, - makeDefElem("verbose", NULL, @2)); + n->options = $2; n->rels = $3; n->is_vacuumcmd = false; $$ = (Node *) n; } - | analyze_keyword '(' utility_option_list ')' opt_vacuum_relation_list + | analyze_keyword VERBOSE opt_vacuum_relation_list { VacuumStmt *n = makeNode(VacuumStmt); - n->options = $3; - n->rels = $5; + n->options = list_make1(makeDefElem("verbose", NULL, @2)); + n->rels = $3; n->is_vacuumcmd = false; $$ = (Node *) n; } ; -utility_option_list: - utility_option_elem - { - $$ = list_make1($1); - } - | utility_option_list ',' utility_option_elem - { - $$ = lappend($1, $3); - } - ; - analyze_keyword: ANALYZE | ANALYSE /* British */ ; -utility_option_elem: - utility_option_name utility_option_arg - { - $$ = makeDefElem($1, $2, @1); - } - ; - -utility_option_name: - NonReservedWord { $$ = $1; } - | analyze_keyword { $$ = "analyze"; } - | FORMAT_LA { $$ = "format"; } - ; - -utility_option_arg: - opt_boolean_or_string { $$ = (Node *) makeString($1); } - | NumericOnly { $$ = (Node *) $1; } - | /* EMPTY */ { $$ = NULL; } - ; - opt_analyze: analyze_keyword { $$ = true; } | /*EMPTY*/ { $$ = false; } diff --git a/src/backend/partitioning/partbounds.c b/src/backend/partitioning/partbounds.c index 4bdc2941efb21..822cf4ec451a4 100644 --- a/src/backend/partitioning/partbounds.c +++ b/src/backend/partitioning/partbounds.c @@ -1007,9 +1007,6 @@ partition_bounds_copy(PartitionBoundInfo src, int ndatums; int nindexes; int partnatts; - bool hash_part; - int natts; - Datum *boundDatums; dest = (PartitionBoundInfo) palloc(sizeof(PartitionBoundInfoData)); @@ -1023,7 +1020,7 @@ partition_bounds_copy(PartitionBoundInfo src, dest->datums = (Datum **) palloc(sizeof(Datum *) * ndatums); - if (src->kind != NULL) + if (src->kind != NULL && ndatums > 0) { PartitionRangeDatumKind *boundKinds; @@ -1058,36 +1055,40 @@ partition_bounds_copy(PartitionBoundInfo src, * For hash partitioning, datums array will have two elements - modulus * and remainder. */ - hash_part = (key->strategy == PARTITION_STRATEGY_HASH); - natts = hash_part ? 2 : partnatts; - boundDatums = palloc(ndatums * natts * sizeof(Datum)); - - for (i = 0; i < ndatums; i++) + if (ndatums > 0) { - int j; - - dest->datums[i] = &boundDatums[i * natts]; + bool hash_part = (key->strategy == PARTITION_STRATEGY_HASH); + int natts = hash_part ? 2 : partnatts; + Datum *boundDatums = palloc(ndatums * natts * sizeof(Datum)); - for (j = 0; j < natts; j++) + for (i = 0; i < ndatums; i++) { - bool byval; - int typlen; + int j; - if (hash_part) - { - typlen = sizeof(int32); /* Always int4 */ - byval = true; /* int4 is pass-by-value */ - } - else + dest->datums[i] = &boundDatums[i * natts]; + + for (j = 0; j < natts; j++) { - byval = key->parttypbyval[j]; - typlen = key->parttyplen[j]; - } + if (dest->kind == NULL || + dest->kind[i][j] == PARTITION_RANGE_DATUM_VALUE) + { + bool byval; + int typlen; - if (dest->kind == NULL || - dest->kind[i][j] == PARTITION_RANGE_DATUM_VALUE) - dest->datums[i][j] = datumCopy(src->datums[i][j], - byval, typlen); + if (hash_part) + { + typlen = sizeof(int32); /* Always int4 */ + byval = true; /* int4 is pass-by-value */ + } + else + { + byval = key->parttypbyval[j]; + typlen = key->parttyplen[j]; + } + dest->datums[i][j] = datumCopy(src->datums[i][j], + byval, typlen); + } + } } } diff --git a/src/backend/postmaster/autovacuum.c b/src/backend/postmaster/autovacuum.c index 9474095f271a1..ff96b36d71025 100644 --- a/src/backend/postmaster/autovacuum.c +++ b/src/backend/postmaster/autovacuum.c @@ -310,6 +310,16 @@ static AutoVacuumShmemStruct *AutoVacuumShmem; static dlist_head DatabaseList = DLIST_STATIC_INIT(DatabaseList); static MemoryContext DatabaseListCxt = NULL; +/* + * Dummy pointer to persuade Valgrind that we've not leaked the array of + * avl_dbase structs. Make it global to ensure the compiler doesn't + * optimize it away. + */ +#ifdef USE_VALGRIND +extern avl_dbase *avl_dbase_array; +avl_dbase *avl_dbase_array; +#endif + /* Pointer to my own WorkerInfo, valid on each worker */ static WorkerInfo MyWorkerInfo = NULL; @@ -562,10 +572,10 @@ AutoVacLauncherMain(const void *startup_data, size_t startup_data_len) /* * Create the initial database list. The invariant we want this list to - * keep is that it's ordered by decreasing next_time. As soon as an entry - * is updated to a higher time, it will be moved to the front (which is - * correct because the only operation is to add autovacuum_naptime to the - * entry, and time always increases). + * keep is that it's ordered by decreasing next_worker. As soon as an + * entry is updated to a higher time, it will be moved to the front (which + * is correct because the only operation is to add autovacuum_naptime to + * the entry, and time always increases). */ rebuild_database_list(InvalidOid); @@ -1020,6 +1030,10 @@ rebuild_database_list(Oid newdb) /* put all the hash elements into an array */ dbary = palloc(nelems * sizeof(avl_dbase)); + /* keep Valgrind quiet */ +#ifdef USE_VALGRIND + avl_dbase_array = dbary; +#endif i = 0; hash_seq_init(&seq, dbhash); @@ -2565,8 +2579,18 @@ do_autovacuum(void) /* * We leak table_toast_map here (among other things), but since we're - * going away soon, it's not a problem. + * going away soon, it's not a problem normally. But when using Valgrind, + * release some stuff to reduce complaints about leaked storage. */ +#ifdef USE_VALGRIND + hash_destroy(table_toast_map); + FreeTupleDesc(pg_class_desc); + if (bstrategy) + pfree(bstrategy); +#endif + + /* Run the rest in xact context, mainly to avoid Valgrind leak warnings */ + MemoryContextSwitchTo(TopTransactionContext); /* * Update pg_database.datfrozenxid, and truncate pg_xact if possible. We diff --git a/src/backend/postmaster/bgworker.c b/src/backend/postmaster/bgworker.c index 116ddf7b835f1..1ad65c237c34e 100644 --- a/src/backend/postmaster/bgworker.c +++ b/src/backend/postmaster/bgworker.c @@ -613,6 +613,7 @@ ResetBackgroundWorkerCrashTimes(void) * resetting. */ rw->rw_crashed_at = 0; + rw->rw_pid = 0; /* * If there was anyone waiting for it, they're history. diff --git a/src/backend/postmaster/checkpointer.c b/src/backend/postmaster/checkpointer.c index 2809e298a44fb..8490148a47d52 100644 --- a/src/backend/postmaster/checkpointer.c +++ b/src/backend/postmaster/checkpointer.c @@ -130,6 +130,13 @@ typedef struct int num_requests; /* current # of requests */ int max_requests; /* allocated array size */ + + int head; /* Index of the first request in the ring + * buffer */ + int tail; /* Index of the last request in the ring + * buffer */ + + /* The ring buffer of pending checkpointer requests */ CheckpointerRequest requests[FLEXIBLE_ARRAY_MEMBER]; } CheckpointerShmemStruct; @@ -138,6 +145,12 @@ static CheckpointerShmemStruct *CheckpointerShmem; /* interval for calling AbsorbSyncRequests in CheckpointWriteDelay */ #define WRITES_PER_ABSORB 1000 +/* Maximum number of checkpointer requests to process in one batch */ +#define CKPT_REQ_BATCH_SIZE 10000 + +/* Max number of requests the checkpointer request queue can hold */ +#define MAX_CHECKPOINT_REQUESTS 10000000 + /* * GUC parameters */ @@ -973,7 +986,8 @@ CheckpointerShmemInit(void) */ MemSet(CheckpointerShmem, 0, size); SpinLockInit(&CheckpointerShmem->ckpt_lck); - CheckpointerShmem->max_requests = NBuffers; + CheckpointerShmem->max_requests = Min(NBuffers, MAX_CHECKPOINT_REQUESTS); + CheckpointerShmem->head = CheckpointerShmem->tail = 0; ConditionVariableInit(&CheckpointerShmem->start_cv); ConditionVariableInit(&CheckpointerShmem->done_cv); } @@ -1201,6 +1215,7 @@ ForwardSyncRequest(const FileTag *ftag, SyncRequestType type) { CheckpointerRequest *request; bool too_full; + int insert_pos; if (!IsUnderPostmaster) return false; /* probably shouldn't even get here */ @@ -1224,10 +1239,14 @@ ForwardSyncRequest(const FileTag *ftag, SyncRequestType type) } /* OK, insert request */ - request = &CheckpointerShmem->requests[CheckpointerShmem->num_requests++]; + insert_pos = CheckpointerShmem->tail; + request = &CheckpointerShmem->requests[insert_pos]; request->ftag = *ftag; request->type = type; + CheckpointerShmem->tail = (CheckpointerShmem->tail + 1) % CheckpointerShmem->max_requests; + CheckpointerShmem->num_requests++; + /* If queue is more than half full, nudge the checkpointer to empty it */ too_full = (CheckpointerShmem->num_requests >= CheckpointerShmem->max_requests / 2); @@ -1269,12 +1288,16 @@ CompactCheckpointerRequestQueue(void) struct CheckpointerSlotMapping { CheckpointerRequest request; - int slot; + int ring_idx; }; - int n, - preserve_count; + int n; int num_skipped = 0; + int head; + int max_requests; + int num_requests; + int read_idx, + write_idx; HASHCTL ctl; HTAB *htab; bool *skip_slot; @@ -1286,8 +1309,13 @@ CompactCheckpointerRequestQueue(void) if (CritSectionCount > 0) return false; + max_requests = CheckpointerShmem->max_requests; + num_requests = CheckpointerShmem->num_requests; + /* Initialize skip_slot array */ - skip_slot = palloc0(sizeof(bool) * CheckpointerShmem->num_requests); + skip_slot = palloc0(sizeof(bool) * max_requests); + + head = CheckpointerShmem->head; /* Initialize temporary hash table */ ctl.keysize = sizeof(CheckpointerRequest); @@ -1311,7 +1339,8 @@ CompactCheckpointerRequestQueue(void) * away preceding entries that would end up being canceled anyhow), but * it's not clear that the extra complexity would buy us anything. */ - for (n = 0; n < CheckpointerShmem->num_requests; n++) + read_idx = head; + for (n = 0; n < num_requests; n++) { CheckpointerRequest *request; struct CheckpointerSlotMapping *slotmap; @@ -1324,16 +1353,19 @@ CompactCheckpointerRequestQueue(void) * CheckpointerShmemInit. Note also that RelFileLocator had better * contain no pad bytes. */ - request = &CheckpointerShmem->requests[n]; + request = &CheckpointerShmem->requests[read_idx]; slotmap = hash_search(htab, request, HASH_ENTER, &found); if (found) { /* Duplicate, so mark the previous occurrence as skippable */ - skip_slot[slotmap->slot] = true; + skip_slot[slotmap->ring_idx] = true; num_skipped++; } /* Remember slot containing latest occurrence of this request value */ - slotmap->slot = n; + slotmap->ring_idx = read_idx; + + /* Move to the next request in the ring buffer */ + read_idx = (read_idx + 1) % max_requests; } /* Done with the hash table. */ @@ -1347,17 +1379,34 @@ CompactCheckpointerRequestQueue(void) } /* We found some duplicates; remove them. */ - preserve_count = 0; - for (n = 0; n < CheckpointerShmem->num_requests; n++) + read_idx = write_idx = head; + for (n = 0; n < num_requests; n++) { - if (skip_slot[n]) - continue; - CheckpointerShmem->requests[preserve_count++] = CheckpointerShmem->requests[n]; + /* If this slot is NOT skipped, keep it */ + if (!skip_slot[read_idx]) + { + /* If the read and write positions are different, copy the request */ + if (write_idx != read_idx) + CheckpointerShmem->requests[write_idx] = + CheckpointerShmem->requests[read_idx]; + + /* Advance the write position */ + write_idx = (write_idx + 1) % max_requests; + } + + read_idx = (read_idx + 1) % max_requests; } + + /* + * Update ring buffer state: head remains the same, tail moves, count + * decreases + */ + CheckpointerShmem->tail = write_idx; + CheckpointerShmem->num_requests -= num_skipped; + ereport(DEBUG1, (errmsg_internal("compacted fsync request queue from %d entries to %d entries", - CheckpointerShmem->num_requests, preserve_count))); - CheckpointerShmem->num_requests = preserve_count; + num_requests, CheckpointerShmem->num_requests))); /* Cleanup. */ pfree(skip_slot); @@ -1378,40 +1427,64 @@ AbsorbSyncRequests(void) { CheckpointerRequest *requests = NULL; CheckpointerRequest *request; - int n; + int n, + i; + bool loop; if (!AmCheckpointerProcess()) return; - LWLockAcquire(CheckpointerCommLock, LW_EXCLUSIVE); - - /* - * We try to avoid holding the lock for a long time by copying the request - * array, and processing the requests after releasing the lock. - * - * Once we have cleared the requests from shared memory, we have to PANIC - * if we then fail to absorb them (eg, because our hashtable runs out of - * memory). This is because the system cannot run safely if we are unable - * to fsync what we have been told to fsync. Fortunately, the hashtable - * is so small that the problem is quite unlikely to arise in practice. - */ - n = CheckpointerShmem->num_requests; - if (n > 0) + do { - requests = (CheckpointerRequest *) palloc(n * sizeof(CheckpointerRequest)); - memcpy(requests, CheckpointerShmem->requests, n * sizeof(CheckpointerRequest)); - } + LWLockAcquire(CheckpointerCommLock, LW_EXCLUSIVE); + + /*--- + * We try to avoid holding the lock for a long time by: + * 1. Copying the request array and processing the requests after + * releasing the lock; + * 2. Processing not the whole queue, but only batches of + * CKPT_REQ_BATCH_SIZE at once. + * + * Once we have cleared the requests from shared memory, we must + * PANIC if we then fail to absorb them (e.g., because our hashtable + * runs out of memory). This is because the system cannot run safely + * if we are unable to fsync what we have been told to fsync. + * Fortunately, the hashtable is so small that the problem is quite + * unlikely to arise in practice. + * + * Note: The maximum possible size of a ring buffer is + * MAX_CHECKPOINT_REQUESTS entries, which fit into a maximum palloc + * allocation size of 1Gb. Our maximum batch size, + * CKPT_REQ_BATCH_SIZE, is even smaller. + */ + n = Min(CheckpointerShmem->num_requests, CKPT_REQ_BATCH_SIZE); + if (n > 0) + { + if (!requests) + requests = (CheckpointerRequest *) palloc(n * sizeof(CheckpointerRequest)); - START_CRIT_SECTION(); + for (i = 0; i < n; i++) + { + requests[i] = CheckpointerShmem->requests[CheckpointerShmem->head]; + CheckpointerShmem->head = (CheckpointerShmem->head + 1) % CheckpointerShmem->max_requests; + } - CheckpointerShmem->num_requests = 0; + CheckpointerShmem->num_requests -= n; - LWLockRelease(CheckpointerCommLock); + } + + START_CRIT_SECTION(); + + /* Are there any requests in the queue? If so, keep going. */ + loop = CheckpointerShmem->num_requests != 0; + + LWLockRelease(CheckpointerCommLock); - for (request = requests; n > 0; request++, n--) - RememberSyncRequest(&request->ftag, request->type); + for (request = requests; n > 0; request++, n--) + RememberSyncRequest(&request->ftag, request->type); - END_CRIT_SECTION(); + END_CRIT_SECTION(); + } while (loop); if (requests) pfree(requests); diff --git a/src/backend/postmaster/pmchild.c b/src/backend/postmaster/pmchild.c index cde1d23a4ca8b..584bb58c8abaf 100644 --- a/src/backend/postmaster/pmchild.c +++ b/src/backend/postmaster/pmchild.c @@ -59,6 +59,17 @@ NON_EXEC_STATIC int num_pmchild_slots = 0; */ dlist_head ActiveChildList; +/* + * Dummy pointer to persuade Valgrind that we've not leaked the array of + * PMChild structs. Make it global to ensure the compiler doesn't + * optimize it away. + */ +#ifdef USE_VALGRIND +extern PMChild *pmchild_array; +PMChild *pmchild_array; +#endif + + /* * MaxLivePostmasterChildren * @@ -125,8 +136,13 @@ InitPostmasterChildSlots(void) for (int i = 0; i < BACKEND_NUM_TYPES; i++) num_pmchild_slots += pmchild_pools[i].size; - /* Initialize them */ + /* Allocate enough slots, and make sure Valgrind doesn't complain */ slots = palloc(num_pmchild_slots * sizeof(PMChild)); +#ifdef USE_VALGRIND + pmchild_array = slots; +#endif + + /* Initialize them */ slotno = 0; for (int btype = 0; btype < BACKEND_NUM_TYPES; btype++) { diff --git a/src/backend/postmaster/postmaster.c b/src/backend/postmaster/postmaster.c index cca9b946e5384..e01d9f0cfe81e 100644 --- a/src/backend/postmaster/postmaster.c +++ b/src/backend/postmaster/postmaster.c @@ -2630,6 +2630,13 @@ CleanupBackend(PMChild *bp, } bp = NULL; + /* + * In a crash case, exit immediately without resetting background worker + * state. However, if restart_after_crash is enabled, the background + * worker state (e.g., rw_pid) still needs be reset so the worker can + * restart after crash recovery. This reset is handled in + * ResetBackgroundWorkerCrashTimes(), not here. + */ if (crashed) { HandleChildCrash(bp_pid, exitstatus, procname); diff --git a/src/backend/replication/libpqwalreceiver/libpqwalreceiver.c b/src/backend/replication/libpqwalreceiver/libpqwalreceiver.c index 886d99951dddf..239641bfbb66a 100644 --- a/src/backend/replication/libpqwalreceiver/libpqwalreceiver.c +++ b/src/backend/replication/libpqwalreceiver/libpqwalreceiver.c @@ -421,31 +421,22 @@ libpqrcv_identify_system(WalReceiverConn *conn, TimeLineID *primary_tli) "IDENTIFY_SYSTEM", WAIT_EVENT_LIBPQWALRECEIVER_RECEIVE); if (PQresultStatus(res) != PGRES_TUPLES_OK) - { - PQclear(res); ereport(ERROR, (errcode(ERRCODE_PROTOCOL_VIOLATION), errmsg("could not receive database system identifier and timeline ID from " "the primary server: %s", pchomp(PQerrorMessage(conn->streamConn))))); - } /* * IDENTIFY_SYSTEM returns 3 columns in 9.3 and earlier, and 4 columns in * 9.4 and onwards. */ if (PQnfields(res) < 3 || PQntuples(res) != 1) - { - int ntuples = PQntuples(res); - int nfields = PQnfields(res); - - PQclear(res); ereport(ERROR, (errcode(ERRCODE_PROTOCOL_VIOLATION), errmsg("invalid response from primary server"), errdetail("Could not identify system: got %d rows and %d fields, expected %d rows and %d or more fields.", - ntuples, nfields, 1, 3))); - } + PQntuples(res), PQnfields(res), 1, 3))); primary_sysid = pstrdup(PQgetvalue(res, 0, 0)); *primary_tli = pg_strtoint32(PQgetvalue(res, 0, 1)); PQclear(res); @@ -607,13 +598,10 @@ libpqrcv_startstreaming(WalReceiverConn *conn, return false; } else if (PQresultStatus(res) != PGRES_COPY_BOTH) - { - PQclear(res); ereport(ERROR, (errcode(ERRCODE_PROTOCOL_VIOLATION), errmsg("could not start WAL streaming: %s", pchomp(PQerrorMessage(conn->streamConn))))); - } PQclear(res); return true; } @@ -721,26 +709,17 @@ libpqrcv_readtimelinehistoryfile(WalReceiverConn *conn, cmd, WAIT_EVENT_LIBPQWALRECEIVER_RECEIVE); if (PQresultStatus(res) != PGRES_TUPLES_OK) - { - PQclear(res); ereport(ERROR, (errcode(ERRCODE_PROTOCOL_VIOLATION), errmsg("could not receive timeline history file from " "the primary server: %s", pchomp(PQerrorMessage(conn->streamConn))))); - } if (PQnfields(res) != 2 || PQntuples(res) != 1) - { - int ntuples = PQntuples(res); - int nfields = PQnfields(res); - - PQclear(res); ereport(ERROR, (errcode(ERRCODE_PROTOCOL_VIOLATION), errmsg("invalid response from primary server"), errdetail("Expected 1 tuple with 2 fields, got %d tuples with %d fields.", - ntuples, nfields))); - } + PQntuples(res), PQnfields(res)))); *filename = pstrdup(PQgetvalue(res, 0, 0)); *len = PQgetlength(res, 0, 1); @@ -844,13 +823,10 @@ libpqrcv_receive(WalReceiverConn *conn, char **buffer, return -1; } else - { - PQclear(res); ereport(ERROR, (errcode(ERRCODE_PROTOCOL_VIOLATION), errmsg("could not receive data from WAL stream: %s", pchomp(PQerrorMessage(conn->streamConn))))); - } } if (rawlen < -1) ereport(ERROR, @@ -974,13 +950,10 @@ libpqrcv_create_slot(WalReceiverConn *conn, const char *slotname, pfree(cmd.data); if (PQresultStatus(res) != PGRES_TUPLES_OK) - { - PQclear(res); ereport(ERROR, (errcode(ERRCODE_PROTOCOL_VIOLATION), errmsg("could not create replication slot \"%s\": %s", slotname, pchomp(PQerrorMessage(conn->streamConn))))); - } if (lsn) *lsn = DatumGetLSN(DirectFunctionCall1Coll(pg_lsn_in, InvalidOid, diff --git a/src/backend/replication/logical/launcher.c b/src/backend/replication/logical/launcher.c index 742d9ba68e900..37377f7eb636c 100644 --- a/src/backend/replication/logical/launcher.c +++ b/src/backend/replication/logical/launcher.c @@ -790,6 +790,8 @@ logicalrep_worker_detach(void) } LWLockRelease(LogicalRepWorkerLock); + + list_free(workers); } /* Block concurrent access. */ diff --git a/src/backend/replication/logical/reorderbuffer.c b/src/backend/replication/logical/reorderbuffer.c index 5febd154b6bae..34cf05668ae84 100644 --- a/src/backend/replication/logical/reorderbuffer.c +++ b/src/backend/replication/logical/reorderbuffer.c @@ -2599,7 +2599,7 @@ ReorderBufferProcessTXN(ReorderBuffer *rb, ReorderBufferTXN *txn, if (++changes_count >= CHANGES_THRESHOLD) { - rb->update_progress_txn(rb, txn, change->lsn); + rb->update_progress_txn(rb, txn, prev_lsn); changes_count = 0; } } diff --git a/src/backend/replication/logical/tablesync.c b/src/backend/replication/logical/tablesync.c index 3fea0a0206ed3..d3356bc84ee0c 100644 --- a/src/backend/replication/logical/tablesync.c +++ b/src/backend/replication/logical/tablesync.c @@ -316,7 +316,8 @@ process_syncing_tables_for_sync(XLogRecPtr current_lsn) UpdateSubscriptionRelState(MyLogicalRepWorker->subid, MyLogicalRepWorker->relid, MyLogicalRepWorker->relstate, - MyLogicalRepWorker->relstate_lsn); + MyLogicalRepWorker->relstate_lsn, + false); /* * End streaming so that LogRepWorkerWalRcvConn can be used to drop @@ -425,6 +426,7 @@ process_syncing_tables_for_apply(XLogRecPtr current_lsn) ListCell *lc; bool started_tx = false; bool should_exit = false; + Relation rel = NULL; Assert(!IsTransactionState()); @@ -492,7 +494,17 @@ process_syncing_tables_for_apply(XLogRecPtr current_lsn) * worker to remove the origin tracking as if there is any * error while dropping we won't restart it to drop the * origin. So passing missing_ok = true. + * + * Lock the subscription and origin in the same order as we + * are doing during DDL commands to avoid deadlocks. See + * AlterSubscription_refresh. */ + LockSharedObject(SubscriptionRelationId, MyLogicalRepWorker->subid, + 0, AccessShareLock); + + if (!rel) + rel = table_open(SubscriptionRelRelationId, RowExclusiveLock); + ReplicationOriginNameForLogicalRep(MyLogicalRepWorker->subid, rstate->relid, originname, @@ -504,7 +516,7 @@ process_syncing_tables_for_apply(XLogRecPtr current_lsn) */ UpdateSubscriptionRelState(MyLogicalRepWorker->subid, rstate->relid, rstate->state, - rstate->lsn); + rstate->lsn, true); } } else @@ -555,7 +567,14 @@ process_syncing_tables_for_apply(XLogRecPtr current_lsn) * This is required to avoid any undetected deadlocks * due to any existing lock as deadlock detector won't * be able to detect the waits on the latch. + * + * Also close any tables prior to the commit. */ + if (rel) + { + table_close(rel, NoLock); + rel = NULL; + } CommitTransactionCommand(); pgstat_report_stat(false); } @@ -623,6 +642,11 @@ process_syncing_tables_for_apply(XLogRecPtr current_lsn) } } + /* Close table if opened */ + if (rel) + table_close(rel, NoLock); + + if (started_tx) { /* @@ -1414,7 +1438,8 @@ LogicalRepSyncTableStart(XLogRecPtr *origin_startpos) UpdateSubscriptionRelState(MyLogicalRepWorker->subid, MyLogicalRepWorker->relid, MyLogicalRepWorker->relstate, - MyLogicalRepWorker->relstate_lsn); + MyLogicalRepWorker->relstate_lsn, + false); CommitTransactionCommand(); pgstat_report_stat(true); @@ -1547,7 +1572,8 @@ LogicalRepSyncTableStart(XLogRecPtr *origin_startpos) UpdateSubscriptionRelState(MyLogicalRepWorker->subid, MyLogicalRepWorker->relid, SUBREL_STATE_FINISHEDCOPY, - MyLogicalRepWorker->relstate_lsn); + MyLogicalRepWorker->relstate_lsn, + false); CommitTransactionCommand(); diff --git a/src/backend/replication/syncrep_scanner.l b/src/backend/replication/syncrep_scanner.l index 7dec1f869c745..02004d621e73d 100644 --- a/src/backend/replication/syncrep_scanner.l +++ b/src/backend/replication/syncrep_scanner.l @@ -157,17 +157,16 @@ syncrep_yyerror(SyncRepConfigData **syncrep_parse_result_p, char **syncrep_parse { struct yyguts_t *yyg = (struct yyguts_t *) yyscanner; /* needed for yytext * macro */ - char *syncrep_parse_error_msg = *syncrep_parse_error_msg_p; /* report only the first error in a parse operation */ - if (syncrep_parse_error_msg) + if (*syncrep_parse_error_msg_p) return; if (yytext[0]) - syncrep_parse_error_msg = psprintf("%s at or near \"%s\"", - message, yytext); + *syncrep_parse_error_msg_p = psprintf("%s at or near \"%s\"", + message, yytext); else - syncrep_parse_error_msg = psprintf("%s at end of input", - message); + *syncrep_parse_error_msg_p = psprintf("%s at end of input", + message); } void diff --git a/src/backend/replication/walsender.c b/src/backend/replication/walsender.c index 4c72a0d43b32b..ee911394a23c6 100644 --- a/src/backend/replication/walsender.c +++ b/src/backend/replication/walsender.c @@ -65,6 +65,7 @@ #include "funcapi.h" #include "libpq/libpq.h" #include "libpq/pqformat.h" +#include "libpq/protocol.h" #include "miscadmin.h" #include "nodes/replnodes.h" #include "pgstat.h" @@ -735,13 +736,13 @@ HandleUploadManifestPacket(StringInfo buf, off_t *offset, switch (mtype) { - case 'd': /* CopyData */ + case PqMsg_CopyData: maxmsglen = PQ_LARGE_MESSAGE_LIMIT; break; - case 'c': /* CopyDone */ - case 'f': /* CopyFail */ - case 'H': /* Flush */ - case 'S': /* Sync */ + case PqMsg_CopyDone: + case PqMsg_CopyFail: + case PqMsg_Flush: + case PqMsg_Sync: maxmsglen = PQ_SMALL_MESSAGE_LIMIT; break; default: @@ -763,19 +764,19 @@ HandleUploadManifestPacket(StringInfo buf, off_t *offset, /* Process the message */ switch (mtype) { - case 'd': /* CopyData */ + case PqMsg_CopyData: AppendIncrementalManifestData(ib, buf->data, buf->len); return true; - case 'c': /* CopyDone */ + case PqMsg_CopyDone: return false; - case 'H': /* Sync */ - case 'S': /* Flush */ + case PqMsg_Sync: + case PqMsg_Flush: /* Ignore these while in CopyOut mode as we do elsewhere. */ return true; - case 'f': + case PqMsg_CopyFail: ereport(ERROR, (errcode(ERRCODE_QUERY_CANCELED), errmsg("COPY from stdin failed: %s", @@ -1569,7 +1570,7 @@ WalSndWriteData(LogicalDecodingContext *ctx, XLogRecPtr lsn, TransactionId xid, tmpbuf.data, sizeof(int64)); /* output previously gathered data in a CopyData packet */ - pq_putmessage_noblock('d', ctx->out->data, ctx->out->len); + pq_putmessage_noblock(PqMsg_CopyData, ctx->out->data, ctx->out->len); CHECK_FOR_INTERRUPTS(); @@ -2305,7 +2306,7 @@ ProcessRepliesIfAny(void) case PqMsg_CopyDone: if (!streamingDoneSending) { - pq_putmessage_noblock('c', NULL, 0); + pq_putmessage_noblock(PqMsg_CopyDone, NULL, 0); streamingDoneSending = true; } @@ -2758,7 +2759,7 @@ ProcessStandbyPSRequestMessage(void) pq_sendint64(&output_message, GetCurrentTimestamp()); /* ... and send it wrapped in CopyData */ - pq_putmessage_noblock('d', output_message.data, output_message.len); + pq_putmessage_noblock(PqMsg_CopyData, output_message.data, output_message.len); } /* @@ -3306,7 +3307,7 @@ XLogSendPhysical(void) wal_segment_close(xlogreader); /* Send CopyDone */ - pq_putmessage_noblock('c', NULL, 0); + pq_putmessage_noblock(PqMsg_CopyDone, NULL, 0); streamingDoneSending = true; WalSndCaughtUp = true; @@ -3434,7 +3435,7 @@ XLogSendPhysical(void) memcpy(&output_message.data[1 + sizeof(int64) + sizeof(int64)], tmpbuf.data, sizeof(int64)); - pq_putmessage_noblock('d', output_message.data, output_message.len); + pq_putmessage_noblock(PqMsg_CopyData, output_message.data, output_message.len); sentPtr = endptr; @@ -4140,7 +4141,7 @@ WalSndKeepalive(bool requestReply, XLogRecPtr writePtr) pq_sendbyte(&output_message, requestReply ? 1 : 0); /* ... and send it wrapped in CopyData */ - pq_putmessage_noblock('d', output_message.data, output_message.len); + pq_putmessage_noblock(PqMsg_CopyData, output_message.data, output_message.len); /* Set local flag */ if (requestReply) diff --git a/src/backend/storage/buffer/bufmgr.c b/src/backend/storage/buffer/bufmgr.c index 6afdd28dba6f2..67431208e7f5f 100644 --- a/src/backend/storage/buffer/bufmgr.c +++ b/src/backend/storage/buffer/bufmgr.c @@ -2743,12 +2743,10 @@ ExtendBufferedRelShared(BufferManagerRelation bmr, * because mdread doesn't complain about reads beyond EOF (when * zero_damaged_pages is ON) and so a previous attempt to read a block * beyond EOF could have left a "valid" zero-filled buffer. - * Unfortunately, we have also seen this case occurring because of - * buggy Linux kernels that sometimes return an lseek(SEEK_END) result - * that doesn't account for a recent write. In that situation, the - * pre-existing buffer would contain valid data that we don't want to - * overwrite. Since the legitimate cases should always have left a - * zero-filled buffer, complain if not PageIsNew. + * + * This has also been observed when relation was overwritten by + * external process. Since the legitimate cases should always have + * left a zero-filled buffer, complain if not PageIsNew. */ if (existing_id >= 0) { @@ -2778,8 +2776,7 @@ ExtendBufferedRelShared(BufferManagerRelation bmr, ereport(ERROR, (errmsg("unexpected data beyond EOF in block %u of relation %s", existing_hdr->tag.blockNum, - relpath(bmr.smgr->smgr_rlocator, fork).str), - errhint("This has been seen to occur with buggy kernels; consider updating your system."))); + relpath(bmr.smgr->smgr_rlocator, fork).str))); /* * We *must* do smgr[zero]extend before succeeding, else the page diff --git a/src/backend/storage/buffer/localbuf.c b/src/backend/storage/buffer/localbuf.c index 3da9c41ee1d7a..3c0d20f4659d2 100644 --- a/src/backend/storage/buffer/localbuf.c +++ b/src/backend/storage/buffer/localbuf.c @@ -932,10 +932,11 @@ GetLocalBufferStorage(void) num_bufs = Min(num_bufs, MaxAllocSize / BLCKSZ); /* Buffers should be I/O aligned. */ - cur_block = (char *) - TYPEALIGN(PG_IO_ALIGN_SIZE, - MemoryContextAlloc(LocalBufferContext, - num_bufs * BLCKSZ + PG_IO_ALIGN_SIZE)); + cur_block = MemoryContextAllocAligned(LocalBufferContext, + num_bufs * BLCKSZ, + PG_IO_ALIGN_SIZE, + 0); + next_buf_in_block = 0; num_bufs_in_block = num_bufs; } diff --git a/src/backend/storage/ipc/latch.c b/src/backend/storage/ipc/latch.c index c6aefd2f688dd..beadeb5e46afa 100644 --- a/src/backend/storage/ipc/latch.c +++ b/src/backend/storage/ipc/latch.c @@ -187,9 +187,11 @@ WaitLatch(Latch *latch, int wakeEvents, long timeout, if (!(wakeEvents & WL_LATCH_SET)) latch = NULL; ModifyWaitEvent(LatchWaitSet, LatchWaitSetLatchPos, WL_LATCH_SET, latch); - ModifyWaitEvent(LatchWaitSet, LatchWaitSetPostmasterDeathPos, - (wakeEvents & (WL_EXIT_ON_PM_DEATH | WL_POSTMASTER_DEATH)), - NULL); + + if (IsUnderPostmaster) + ModifyWaitEvent(LatchWaitSet, LatchWaitSetPostmasterDeathPos, + (wakeEvents & (WL_EXIT_ON_PM_DEATH | WL_POSTMASTER_DEATH)), + NULL); if (WaitEventSetWait(LatchWaitSet, (wakeEvents & WL_TIMEOUT) ? timeout : -1, diff --git a/src/backend/storage/ipc/procsignal.c b/src/backend/storage/ipc/procsignal.c index a9bb540b55ac2..087821311cceb 100644 --- a/src/backend/storage/ipc/procsignal.c +++ b/src/backend/storage/ipc/procsignal.c @@ -728,7 +728,11 @@ procsignal_sigusr1_handler(SIGNAL_ARGS) void SendCancelRequest(int backendPID, const uint8 *cancel_key, int cancel_key_len) { - Assert(backendPID != 0); + if (backendPID == 0) + { + ereport(LOG, (errmsg("invalid cancel request with PID 0"))); + return; + } /* * See if we have a matching backend. Reading the pss_pid and diff --git a/src/backend/storage/lmgr/generate-lwlocknames.pl b/src/backend/storage/lmgr/generate-lwlocknames.pl index c7a6720440db6..cd3e43c448aed 100644 --- a/src/backend/storage/lmgr/generate-lwlocknames.pl +++ b/src/backend/storage/lmgr/generate-lwlocknames.pl @@ -27,18 +27,24 @@ # -# First, record the predefined LWLocks listed in wait_event_names.txt. We'll -# cross-check those with the ones in lwlocklist.h. +# First, record the predefined LWLocks and built-in tranches listed in +# wait_event_names.txt. We'll cross-check those with the ones in lwlocklist.h. # +my @wait_event_tranches; my @wait_event_lwlocks; my $record_lwlocks = 0; +my $in_tranches = 0; while (<$wait_event_names>) { chomp; # Check for end marker. - last if /^# END OF PREDEFINED LWLOCKS/; + if (/^# END OF PREDEFINED LWLOCKS/) + { + $in_tranches = 1; + next; + } # Skip comments and empty lines. next if /^#/; @@ -54,13 +60,29 @@ # Go to the next line if we are not yet recording LWLocks. next if not $record_lwlocks; + # Stop recording if we reach another section. + last if /^Section:/; + # Record the LWLock. (my $waiteventname, my $waitevendocsentence) = split(/\t/, $_); - push(@wait_event_lwlocks, $waiteventname); + + if ($in_tranches) + { + push(@wait_event_tranches, $waiteventname); + } + else + { + push(@wait_event_lwlocks, $waiteventname); + } } +# +# While gathering the list of predefined LWLocks, cross-check the lists in +# lwlocklist.h with the wait events we just recorded. +# my $in_comment = 0; -my $i = 0; +my $lwlock_count = 0; +my $tranche_count = 0; while (<$lwlocklist>) { chomp; @@ -81,38 +103,72 @@ next; } - die "unable to parse lwlocklist.h line \"$_\"" - unless /^PG_LWLOCK\((\d+),\s+(\w+)\)$/; + # + # Gather list of predefined LWLocks and cross-check with the wait events. + # + if (/^PG_LWLOCK\((\d+),\s+(\w+)\)$/) + { + my ($lockidx, $lockname) = ($1, $2); - (my $lockidx, my $lockname) = ($1, $2); + die "lwlocklist.h not in order" if $lockidx < $lastlockidx; + die "lwlocklist.h has duplicates" if $lockidx == $lastlockidx; - die "lwlocklist.h not in order" if $lockidx < $lastlockidx; - die "lwlocklist.h has duplicates" if $lockidx == $lastlockidx; + die "$lockname defined in lwlocklist.h but missing from " + . "wait_event_names.txt" + if $lwlock_count >= scalar @wait_event_lwlocks; + die "lists of predefined LWLocks do not match (first mismatch at " + . "$wait_event_lwlocks[$lwlock_count] in wait_event_names.txt and " + . "$lockname in lwlocklist.h)" + if $wait_event_lwlocks[$lwlock_count] ne $lockname; - die "$lockname defined in lwlocklist.h but missing from " - . "wait_event_names.txt" - if $i >= scalar @wait_event_lwlocks; - die "lists of predefined LWLocks do not match (first mismatch at " - . "$wait_event_lwlocks[$i] in wait_event_names.txt and $lockname in " - . "lwlocklist.h)" - if $wait_event_lwlocks[$i] ne $lockname; - $i++; + $lwlock_count++; - while ($lastlockidx < $lockidx - 1) + while ($lastlockidx < $lockidx - 1) + { + ++$lastlockidx; + } + $lastlockidx = $lockidx; + + # Add a "Lock" suffix to each lock name, as the C code depends on that. + printf $h "#define %-32s (&MainLWLockArray[$lockidx].lock)\n", + $lockname . "Lock"; + + next; + } + + # + # Cross-check the built-in LWLock tranches with the wait events. + # + if (/^PG_LWLOCKTRANCHE\((\w+),\s+(\w+)\)$/) { - ++$lastlockidx; + my ($tranche_id, $tranche_name) = ($1, $2); + + die "$tranche_name defined in lwlocklist.h but missing from " + . "wait_event_names.txt" + if $tranche_count >= scalar @wait_event_tranches; + die + "lists of built-in LWLock tranches do not match (first mismatch at " + . "$wait_event_tranches[$tranche_count] in wait_event_names.txt and " + . "$tranche_name in lwlocklist.h)" + if $wait_event_tranches[$tranche_count] ne $tranche_name; + + $tranche_count++; + + next; } - $lastlockidx = $lockidx; - # Add a "Lock" suffix to each lock name, as the C code depends on that - printf $h "#define %-32s (&MainLWLockArray[$lockidx].lock)\n", - $lockname . "Lock"; + die "unable to parse lwlocklist.h line \"$_\""; } die - "$wait_event_lwlocks[$i] defined in wait_event_names.txt but missing from " - . "lwlocklist.h" - if $i < scalar @wait_event_lwlocks; + "$wait_event_lwlocks[$lwlock_count] defined in wait_event_names.txt but " + . " missing from lwlocklist.h" + if $lwlock_count < scalar @wait_event_lwlocks; + +die + "$wait_event_tranches[$tranche_count] defined in wait_event_names.txt but " + . "missing from lwlocklist.h" + if $tranche_count < scalar @wait_event_tranches; print $h "\n"; printf $h "#define NUM_INDIVIDUAL_LWLOCKS %s\n", $lastlockidx + 1; diff --git a/src/backend/storage/lmgr/lwlock.c b/src/backend/storage/lmgr/lwlock.c index 2d43bf2cc1323..ec9c345ffdfb8 100644 --- a/src/backend/storage/lmgr/lwlock.c +++ b/src/backend/storage/lmgr/lwlock.c @@ -122,9 +122,8 @@ StaticAssertDecl((LW_VAL_EXCLUSIVE & LW_FLAG_MASK) == 0, * own tranche. We absorb the names of these tranches from there into * BuiltinTrancheNames here. * - * 2. There are some predefined tranches for built-in groups of locks. - * These are listed in enum BuiltinTrancheIds in lwlock.h, and their names - * appear in BuiltinTrancheNames[] below. + * 2. There are some predefined tranches for built-in groups of locks defined + * in lwlocklist.h. We absorb the names of these tranches, too. * * 3. Extensions can create new tranches, via either RequestNamedLWLockTranche * or LWLockRegisterTranche. The names of these that are known in the current @@ -135,49 +134,10 @@ StaticAssertDecl((LW_VAL_EXCLUSIVE & LW_FLAG_MASK) == 0, */ static const char *const BuiltinTrancheNames[] = { #define PG_LWLOCK(id, lockname) [id] = CppAsString(lockname), +#define PG_LWLOCKTRANCHE(id, lockname) [LWTRANCHE_##id] = CppAsString(lockname), #include "storage/lwlocklist.h" #undef PG_LWLOCK - [LWTRANCHE_XACT_BUFFER] = "XactBuffer", - [LWTRANCHE_COMMITTS_BUFFER] = "CommitTsBuffer", - [LWTRANCHE_SUBTRANS_BUFFER] = "SubtransBuffer", - [LWTRANCHE_MULTIXACTOFFSET_BUFFER] = "MultiXactOffsetBuffer", - [LWTRANCHE_MULTIXACTMEMBER_BUFFER] = "MultiXactMemberBuffer", - [LWTRANCHE_NOTIFY_BUFFER] = "NotifyBuffer", - [LWTRANCHE_SERIAL_BUFFER] = "SerialBuffer", - [LWTRANCHE_WAL_INSERT] = "WALInsert", - [LWTRANCHE_BUFFER_CONTENT] = "BufferContent", - [LWTRANCHE_REPLICATION_ORIGIN_STATE] = "ReplicationOriginState", - [LWTRANCHE_REPLICATION_SLOT_IO] = "ReplicationSlotIO", - [LWTRANCHE_LOCK_FASTPATH] = "LockFastPath", - [LWTRANCHE_BUFFER_MAPPING] = "BufferMapping", - [LWTRANCHE_LOCK_MANAGER] = "LockManager", - [LWTRANCHE_PREDICATE_LOCK_MANAGER] = "PredicateLockManager", - [LWTRANCHE_PARALLEL_HASH_JOIN] = "ParallelHashJoin", - [LWTRANCHE_PARALLEL_BTREE_SCAN] = "ParallelBtreeScan", - [LWTRANCHE_PARALLEL_QUERY_DSA] = "ParallelQueryDSA", - [LWTRANCHE_PER_SESSION_DSA] = "PerSessionDSA", - [LWTRANCHE_PER_SESSION_RECORD_TYPE] = "PerSessionRecordType", - [LWTRANCHE_PER_SESSION_RECORD_TYPMOD] = "PerSessionRecordTypmod", - [LWTRANCHE_SHARED_TUPLESTORE] = "SharedTupleStore", - [LWTRANCHE_SHARED_TIDBITMAP] = "SharedTidBitmap", - [LWTRANCHE_PARALLEL_APPEND] = "ParallelAppend", - [LWTRANCHE_PER_XACT_PREDICATE_LIST] = "PerXactPredicateList", - [LWTRANCHE_PGSTATS_DSA] = "PgStatsDSA", - [LWTRANCHE_PGSTATS_HASH] = "PgStatsHash", - [LWTRANCHE_PGSTATS_DATA] = "PgStatsData", - [LWTRANCHE_LAUNCHER_DSA] = "LogicalRepLauncherDSA", - [LWTRANCHE_LAUNCHER_HASH] = "LogicalRepLauncherHash", - [LWTRANCHE_DSM_REGISTRY_DSA] = "DSMRegistryDSA", - [LWTRANCHE_DSM_REGISTRY_HASH] = "DSMRegistryHash", - [LWTRANCHE_COMMITTS_SLRU] = "CommitTsSLRU", - [LWTRANCHE_MULTIXACTOFFSET_SLRU] = "MultiXactOffsetSLRU", - [LWTRANCHE_MULTIXACTMEMBER_SLRU] = "MultiXactMemberSLRU", - [LWTRANCHE_NOTIFY_SLRU] = "NotifySLRU", - [LWTRANCHE_SERIAL_SLRU] = "SerialSLRU", - [LWTRANCHE_SUBTRANS_SLRU] = "SubtransSLRU", - [LWTRANCHE_XACT_SLRU] = "XactSLRU", - [LWTRANCHE_PARALLEL_VACUUM_DSA] = "ParallelVacuumDSA", - [LWTRANCHE_AIO_URING_COMPLETION] = "AioUringCompletion", +#undef PG_LWLOCKTRANCHE }; StaticAssertDecl(lengthof(BuiltinTrancheNames) == diff --git a/src/backend/tcop/backend_startup.c b/src/backend/tcop/backend_startup.c index ad0af5edc1f21..14d5fc0b1965a 100644 --- a/src/backend/tcop/backend_startup.c +++ b/src/backend/tcop/backend_startup.c @@ -492,7 +492,7 @@ static int ProcessStartupPacket(Port *port, bool ssl_done, bool gss_done) { int32 len; - char *buf; + char *buf = NULL; ProtocolVersion proto; MemoryContext oldcontext; @@ -516,7 +516,7 @@ ProcessStartupPacket(Port *port, bool ssl_done, bool gss_done) * scanners, which may be less benign, but it's not really our job to * notice those.) */ - return STATUS_ERROR; + goto fail; } if (pq_getbytes(((char *) &len) + 1, 3) == EOF) @@ -526,7 +526,7 @@ ProcessStartupPacket(Port *port, bool ssl_done, bool gss_done) ereport(COMMERROR, (errcode(ERRCODE_PROTOCOL_VIOLATION), errmsg("incomplete startup packet"))); - return STATUS_ERROR; + goto fail; } len = pg_ntoh32(len); @@ -538,7 +538,7 @@ ProcessStartupPacket(Port *port, bool ssl_done, bool gss_done) ereport(COMMERROR, (errcode(ERRCODE_PROTOCOL_VIOLATION), errmsg("invalid length of startup packet"))); - return STATUS_ERROR; + goto fail; } /* @@ -554,7 +554,7 @@ ProcessStartupPacket(Port *port, bool ssl_done, bool gss_done) ereport(COMMERROR, (errcode(ERRCODE_PROTOCOL_VIOLATION), errmsg("incomplete startup packet"))); - return STATUS_ERROR; + goto fail; } pq_endmsgread(); @@ -568,7 +568,7 @@ ProcessStartupPacket(Port *port, bool ssl_done, bool gss_done) { ProcessCancelRequestPacket(port, buf, len); /* Not really an error, but we don't want to proceed further */ - return STATUS_ERROR; + goto fail; } if (proto == NEGOTIATE_SSL_CODE && !ssl_done) @@ -607,14 +607,16 @@ ProcessStartupPacket(Port *port, bool ssl_done, bool gss_done) ereport(COMMERROR, (errcode_for_socket_access(), errmsg("failed to send SSL negotiation response: %m"))); - return STATUS_ERROR; /* close the connection */ + goto fail; /* close the connection */ } #ifdef USE_SSL if (SSLok == 'S' && secure_open_server(port) == -1) - return STATUS_ERROR; + goto fail; #endif + pfree(buf); + /* * At this point we should have no data already buffered. If we do, * it was received before we performed the SSL handshake, so it wasn't @@ -661,14 +663,16 @@ ProcessStartupPacket(Port *port, bool ssl_done, bool gss_done) ereport(COMMERROR, (errcode_for_socket_access(), errmsg("failed to send GSSAPI negotiation response: %m"))); - return STATUS_ERROR; /* close the connection */ + goto fail; /* close the connection */ } #ifdef ENABLE_GSS if (GSSok == 'G' && secure_open_gssapi(port) == -1) - return STATUS_ERROR; + goto fail; #endif + pfree(buf); + /* * At this point we should have no data already buffered. If we do, * it was received before we performed the GSS handshake, so it wasn't @@ -863,7 +867,16 @@ ProcessStartupPacket(Port *port, bool ssl_done, bool gss_done) */ MemoryContextSwitchTo(oldcontext); + pfree(buf); + return STATUS_OK; + +fail: + /* be tidy, just to avoid Valgrind complaints */ + if (buf) + pfree(buf); + + return STATUS_ERROR; } /* diff --git a/src/backend/tcop/postgres.c b/src/backend/tcop/postgres.c index 2f8c3d5f91822..0cecd4649020f 100644 --- a/src/backend/tcop/postgres.c +++ b/src/backend/tcop/postgres.c @@ -988,6 +988,7 @@ pg_plan_queries(List *querytrees, const char *query_string, int cursorOptions, stmt->stmt_location = query->stmt_location; stmt->stmt_len = query->stmt_len; stmt->queryId = query->queryId; + stmt->planOrigin = PLAN_STMT_INTERNAL; } else { diff --git a/src/backend/tcop/utility.c b/src/backend/tcop/utility.c index 4c1faf5575c4d..4f4191b0ea6b4 100644 --- a/src/backend/tcop/utility.c +++ b/src/backend/tcop/utility.c @@ -1234,6 +1234,7 @@ ProcessUtilitySlow(ParseState *pstate, wrapper->utilityStmt = stmt; wrapper->stmt_location = pstmt->stmt_location; wrapper->stmt_len = pstmt->stmt_len; + wrapper->planOrigin = PLAN_STMT_INTERNAL; ProcessUtility(wrapper, queryString, @@ -1964,6 +1965,7 @@ ProcessUtilityForAlterTable(Node *stmt, AlterTableUtilityContext *context) wrapper->utilityStmt = stmt; wrapper->stmt_location = context->pstmt->stmt_location; wrapper->stmt_len = context->pstmt->stmt_len; + wrapper->planOrigin = PLAN_STMT_INTERNAL; ProcessUtility(wrapper, context->queryString, diff --git a/src/backend/tsearch/dict_ispell.c b/src/backend/tsearch/dict_ispell.c index 63bd193a78a89..debfbf956cc1f 100644 --- a/src/backend/tsearch/dict_ispell.c +++ b/src/backend/tsearch/dict_ispell.c @@ -47,24 +47,30 @@ dispell_init(PG_FUNCTION_ARGS) if (strcmp(defel->defname, "dictfile") == 0) { + char *filename; + if (dictloaded) ereport(ERROR, (errcode(ERRCODE_INVALID_PARAMETER_VALUE), errmsg("multiple DictFile parameters"))); - NIImportDictionary(&(d->obj), - get_tsearch_config_filename(defGetString(defel), - "dict")); + filename = get_tsearch_config_filename(defGetString(defel), + "dict"); + NIImportDictionary(&(d->obj), filename); + pfree(filename); dictloaded = true; } else if (strcmp(defel->defname, "afffile") == 0) { + char *filename; + if (affloaded) ereport(ERROR, (errcode(ERRCODE_INVALID_PARAMETER_VALUE), errmsg("multiple AffFile parameters"))); - NIImportAffixes(&(d->obj), - get_tsearch_config_filename(defGetString(defel), - "affix")); + filename = get_tsearch_config_filename(defGetString(defel), + "affix"); + NIImportAffixes(&(d->obj), filename); + pfree(filename); affloaded = true; } else if (strcmp(defel->defname, "stopwords") == 0) diff --git a/src/backend/tsearch/dict_synonym.c b/src/backend/tsearch/dict_synonym.c index 0da5a9d686802..c2773eb01adee 100644 --- a/src/backend/tsearch/dict_synonym.c +++ b/src/backend/tsearch/dict_synonym.c @@ -199,6 +199,7 @@ dsynonym_init(PG_FUNCTION_ARGS) } tsearch_readline_end(&trst); + pfree(filename); d->len = cur; qsort(d->syn, d->len, sizeof(Syn), compareSyn); diff --git a/src/backend/tsearch/dict_thesaurus.c b/src/backend/tsearch/dict_thesaurus.c index 1bebe36a6910e..1e6bbde1ca7d8 100644 --- a/src/backend/tsearch/dict_thesaurus.c +++ b/src/backend/tsearch/dict_thesaurus.c @@ -167,17 +167,17 @@ addWrd(DictThesaurus *d, char *b, char *e, uint32 idsubst, uint16 nwrd, uint16 p static void thesaurusRead(const char *filename, DictThesaurus *d) { + char *real_filename = get_tsearch_config_filename(filename, "ths"); tsearch_readline_state trst; uint32 idsubst = 0; bool useasis = false; char *line; - filename = get_tsearch_config_filename(filename, "ths"); - if (!tsearch_readline_begin(&trst, filename)) + if (!tsearch_readline_begin(&trst, real_filename)) ereport(ERROR, (errcode(ERRCODE_CONFIG_FILE_ERROR), errmsg("could not open thesaurus file \"%s\": %m", - filename))); + real_filename))); while ((line = tsearch_readline(&trst)) != NULL) { @@ -297,6 +297,7 @@ thesaurusRead(const char *filename, DictThesaurus *d) d->nsubst = idsubst; tsearch_readline_end(&trst); + pfree(real_filename); } static TheLexeme * diff --git a/src/backend/utils/activity/pgstat.c b/src/backend/utils/activity/pgstat.c index 8b57845e8709f..6bc91ce0dadda 100644 --- a/src/backend/utils/activity/pgstat.c +++ b/src/backend/utils/activity/pgstat.c @@ -212,6 +212,11 @@ int pgstat_fetch_consistency = PGSTAT_FETCH_CONSISTENCY_CACHE; PgStat_LocalState pgStatLocal; +/* + * Track pending reports for fixed-numbered stats, used by + * pgstat_report_stat(). + */ +bool pgstat_report_fixed = false; /* ---------- * Local data @@ -370,7 +375,6 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE] .shared_data_off = offsetof(PgStatShared_Backend, stats), .shared_data_len = sizeof(((PgStatShared_Backend *) 0)->stats), - .have_static_pending_cb = pgstat_backend_have_pending_cb, .flush_static_cb = pgstat_backend_flush_cb, .reset_timestamp_cb = pgstat_backend_reset_timestamp_cb, }, @@ -437,7 +441,6 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE] .shared_data_len = sizeof(((PgStatShared_IO *) 0)->stats), .flush_static_cb = pgstat_io_flush_cb, - .have_static_pending_cb = pgstat_io_have_pending_cb, .init_shmem_cb = pgstat_io_init_shmem_cb, .reset_all_cb = pgstat_io_reset_all_cb, .snapshot_cb = pgstat_io_snapshot_cb, @@ -455,7 +458,6 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE] .shared_data_len = sizeof(((PgStatShared_SLRU *) 0)->stats), .flush_static_cb = pgstat_slru_flush_cb, - .have_static_pending_cb = pgstat_slru_have_pending_cb, .init_shmem_cb = pgstat_slru_init_shmem_cb, .reset_all_cb = pgstat_slru_reset_all_cb, .snapshot_cb = pgstat_slru_snapshot_cb, @@ -474,7 +476,6 @@ static const PgStat_KindInfo pgstat_kind_builtin_infos[PGSTAT_KIND_BUILTIN_SIZE] .init_backend_cb = pgstat_wal_init_backend_cb, .flush_static_cb = pgstat_wal_flush_cb, - .have_static_pending_cb = pgstat_wal_have_pending_cb, .init_shmem_cb = pgstat_wal_init_shmem_cb, .reset_all_cb = pgstat_wal_reset_all_cb, .snapshot_cb = pgstat_wal_snapshot_cb, @@ -708,29 +709,10 @@ pgstat_report_stat(bool force) } /* Don't expend a clock check if nothing to do */ - if (dlist_is_empty(&pgStatPending)) + if (dlist_is_empty(&pgStatPending) && + !pgstat_report_fixed) { - bool do_flush = false; - - /* Check for pending stats */ - for (PgStat_Kind kind = PGSTAT_KIND_MIN; kind <= PGSTAT_KIND_MAX; kind++) - { - const PgStat_KindInfo *kind_info = pgstat_get_kind_info(kind); - - if (!kind_info) - continue; - if (!kind_info->have_static_pending_cb) - continue; - - if (kind_info->have_static_pending_cb()) - { - do_flush = true; - break; - } - } - - if (!do_flush) - return 0; + return 0; } /* @@ -784,16 +766,19 @@ pgstat_report_stat(bool force) partial_flush |= pgstat_flush_pending_entries(nowait); /* flush of other stats kinds */ - for (PgStat_Kind kind = PGSTAT_KIND_MIN; kind <= PGSTAT_KIND_MAX; kind++) + if (pgstat_report_fixed) { - const PgStat_KindInfo *kind_info = pgstat_get_kind_info(kind); + for (PgStat_Kind kind = PGSTAT_KIND_MIN; kind <= PGSTAT_KIND_MAX; kind++) + { + const PgStat_KindInfo *kind_info = pgstat_get_kind_info(kind); - if (!kind_info) - continue; - if (!kind_info->flush_static_cb) - continue; + if (!kind_info) + continue; + if (!kind_info->flush_static_cb) + continue; - partial_flush |= kind_info->flush_static_cb(nowait); + partial_flush |= kind_info->flush_static_cb(nowait); + } } last_flush = now; @@ -815,6 +800,7 @@ pgstat_report_stat(bool force) } pending_since = 0; + pgstat_report_fixed = false; return 0; } diff --git a/src/backend/utils/activity/pgstat_backend.c b/src/backend/utils/activity/pgstat_backend.c index 51256277e8d37..8714a85e2d936 100644 --- a/src/backend/utils/activity/pgstat_backend.c +++ b/src/backend/utils/activity/pgstat_backend.c @@ -66,6 +66,7 @@ pgstat_count_backend_io_op_time(IOObject io_object, IOContext io_context, io_time); backend_has_iostats = true; + pgstat_report_fixed = true; } void @@ -81,6 +82,7 @@ pgstat_count_backend_io_op(IOObject io_object, IOContext io_context, PendingBackendStats.pending_io.bytes[io_object][io_context][io_op] += bytes; backend_has_iostats = true; + pgstat_report_fixed = true; } /* @@ -301,18 +303,6 @@ pgstat_flush_backend(bool nowait, bits32 flags) return false; } -/* - * Check if there are any backend stats waiting for flush. - */ -bool -pgstat_backend_have_pending_cb(void) -{ - if (!pgstat_tracks_backend_bktype(MyBackendType)) - return false; - - return (backend_has_iostats || pgstat_backend_wal_have_pending()); -} - /* * Callback to flush out locally pending backend statistics. * diff --git a/src/backend/utils/activity/pgstat_io.c b/src/backend/utils/activity/pgstat_io.c index d8d26379a571e..13ae57ed6498d 100644 --- a/src/backend/utils/activity/pgstat_io.c +++ b/src/backend/utils/activity/pgstat_io.c @@ -80,6 +80,7 @@ pgstat_count_io_op(IOObject io_object, IOContext io_context, IOOp io_op, pgstat_count_backend_io_op(io_object, io_context, io_op, cnt, bytes); have_iostats = true; + pgstat_report_fixed = true; } /* @@ -167,15 +168,6 @@ pgstat_fetch_stat_io(void) return &pgStatLocal.snapshot.io; } -/* - * Check if there any IO stats waiting for flush. - */ -bool -pgstat_io_have_pending_cb(void) -{ - return have_iostats; -} - /* * Simpler wrapper of pgstat_io_flush_cb() */ diff --git a/src/backend/utils/activity/pgstat_slru.c b/src/backend/utils/activity/pgstat_slru.c index b9e940dde45b6..7bd8744accb0e 100644 --- a/src/backend/utils/activity/pgstat_slru.c +++ b/src/backend/utils/activity/pgstat_slru.c @@ -143,15 +143,6 @@ pgstat_get_slru_index(const char *name) return (SLRU_NUM_ELEMENTS - 1); } -/* - * Check if there are any SLRU stats entries waiting for flush. - */ -bool -pgstat_slru_have_pending_cb(void) -{ - return have_slrustats; -} - /* * Flush out locally pending SLRU stats entries * @@ -247,6 +238,7 @@ get_slru_entry(int slru_idx) Assert((slru_idx >= 0) && (slru_idx < SLRU_NUM_ELEMENTS)); have_slrustats = true; + pgstat_report_fixed = true; return &pending_SLRUStats[slru_idx]; } diff --git a/src/backend/utils/activity/pgstat_wal.c b/src/backend/utils/activity/pgstat_wal.c index 16a1ecb4d90d2..0d04480d2f6d0 100644 --- a/src/backend/utils/activity/pgstat_wal.c +++ b/src/backend/utils/activity/pgstat_wal.c @@ -71,6 +71,15 @@ pgstat_fetch_stat_wal(void) return &pgStatLocal.snapshot.wal; } +/* + * To determine whether WAL usage happened. + */ +static inline bool +pgstat_wal_have_pending(void) +{ + return pgWalUsage.wal_records != prevWalUsage.wal_records; +} + /* * Calculate how much WAL usage counters have increased by subtracting the * previous counters from the current ones. @@ -92,7 +101,7 @@ pgstat_wal_flush_cb(bool nowait) * This function can be called even if nothing at all has happened. Avoid * taking lock for nothing in that case. */ - if (!pgstat_wal_have_pending_cb()) + if (!pgstat_wal_have_pending()) return false; /* @@ -136,15 +145,6 @@ pgstat_wal_init_backend_cb(void) prevWalUsage = pgWalUsage; } -/* - * To determine whether WAL usage happened. - */ -bool -pgstat_wal_have_pending_cb(void) -{ - return pgWalUsage.wal_records != prevWalUsage.wal_records; -} - void pgstat_wal_init_shmem_cb(void *stats) { diff --git a/src/backend/utils/activity/wait_event_names.txt b/src/backend/utils/activity/wait_event_names.txt index 4da68312b5f97..0be307d2ca04b 100644 --- a/src/backend/utils/activity/wait_event_names.txt +++ b/src/backend/utils/activity/wait_event_names.txt @@ -356,9 +356,13 @@ AioWorkerSubmissionQueue "Waiting to access AIO worker submission queue." # # END OF PREDEFINED LWLOCKS (DO NOT CHANGE THIS LINE) # -# Predefined LWLocks (i.e., those declared in lwlocknames.h) must be listed -# in the section above and must be listed in the same order as in -# lwlocknames.h. Other LWLocks must be listed in the section below. +# Predefined LWLocks (i.e., those declared at the top of lwlocknames.h) must be +# listed in the section above and must be listed in the same order as in +# lwlocknames.h. +# +# Likewise, the built-in LWLock tranches (i.e., those declared at the bottom of +# lwlocknames.h) must be listed in the section below and must be listed in the +# same order as in lwlocknames.h. # XactBuffer "Waiting for I/O on a transaction status SLRU buffer." diff --git a/src/backend/utils/adt/tid.c b/src/backend/utils/adt/tid.c index 1b0df1117171a..39dab3e42df58 100644 --- a/src/backend/utils/adt/tid.c +++ b/src/backend/utils/adt/tid.c @@ -84,7 +84,7 @@ tidin(PG_FUNCTION_ARGS) /* * Cope with possibility that unsigned long is wider than BlockNumber, in * which case strtoul will not raise an error for some values that are out - * of the range of BlockNumber. (See similar code in oidin().) + * of the range of BlockNumber. (See similar code in uint32in_subr().) */ #if SIZEOF_LONG > 4 if (cvt != (unsigned long) blockNumber && diff --git a/src/backend/utils/adt/xml.c b/src/backend/utils/adt/xml.c index f7b731825fca0..182e8f75db75c 100644 --- a/src/backend/utils/adt/xml.c +++ b/src/backend/utils/adt/xml.c @@ -1769,7 +1769,7 @@ xml_doctype_in_content(const xmlChar *str) * xmloption_arg, but a DOCTYPE node in the input can force DOCUMENT mode). * * If parsed_nodes isn't NULL and we parse in CONTENT mode, the list - * of parsed nodes from the xmlParseInNodeContext call will be returned + * of parsed nodes from the xmlParseBalancedChunkMemory call will be returned * to *parsed_nodes. (It is caller's responsibility to free that.) * * Errors normally result in ereport(ERROR), but if escontext is an @@ -1795,6 +1795,7 @@ xml_parse(text *data, XmlOptionType xmloption_arg, PgXmlErrorContext *xmlerrcxt; volatile xmlParserCtxtPtr ctxt = NULL; volatile xmlDocPtr doc = NULL; + volatile int save_keep_blanks = -1; /* * This step looks annoyingly redundant, but we must do it to have a @@ -1822,7 +1823,6 @@ xml_parse(text *data, XmlOptionType xmloption_arg, PG_TRY(); { bool parse_as_document = false; - int options; int res_code; size_t count = 0; xmlChar *version = NULL; @@ -1853,18 +1853,6 @@ xml_parse(text *data, XmlOptionType xmloption_arg, parse_as_document = true; } - /* - * Select parse options. - * - * Note that here we try to apply DTD defaults (XML_PARSE_DTDATTR) - * according to SQL/XML:2008 GR 10.16.7.d: 'Default values defined by - * internal DTD are applied'. As for external DTDs, we try to support - * them too (see SQL/XML:2008 GR 10.16.7.e), but that doesn't really - * happen because xmlPgEntityLoader prevents it. - */ - options = XML_PARSE_NOENT | XML_PARSE_DTDATTR - | (preserve_whitespace ? 0 : XML_PARSE_NOBLANKS); - /* initialize output parameters */ if (parsed_xmloptiontype != NULL) *parsed_xmloptiontype = parse_as_document ? XMLOPTION_DOCUMENT : @@ -1874,11 +1862,26 @@ xml_parse(text *data, XmlOptionType xmloption_arg, if (parse_as_document) { + int options; + + /* set up parser context used by xmlCtxtReadDoc */ ctxt = xmlNewParserCtxt(); if (ctxt == NULL || xmlerrcxt->err_occurred) xml_ereport(xmlerrcxt, ERROR, ERRCODE_OUT_OF_MEMORY, "could not allocate parser context"); + /* + * Select parse options. + * + * Note that here we try to apply DTD defaults (XML_PARSE_DTDATTR) + * according to SQL/XML:2008 GR 10.16.7.d: 'Default values defined + * by internal DTD are applied'. As for external DTDs, we try to + * support them too (see SQL/XML:2008 GR 10.16.7.e), but that + * doesn't really happen because xmlPgEntityLoader prevents it. + */ + options = XML_PARSE_NOENT | XML_PARSE_DTDATTR + | (preserve_whitespace ? 0 : XML_PARSE_NOBLANKS); + doc = xmlCtxtReadDoc(ctxt, utf8string, NULL, /* no URL */ "UTF-8", @@ -1900,10 +1903,7 @@ xml_parse(text *data, XmlOptionType xmloption_arg, } else { - xmlNodePtr root; - xmlNodePtr oldroot PG_USED_FOR_ASSERTS_ONLY; - - /* set up document with empty root node to be the context node */ + /* set up document that xmlParseBalancedChunkMemory will add to */ doc = xmlNewDoc(version); if (doc == NULL || xmlerrcxt->err_occurred) xml_ereport(xmlerrcxt, ERROR, ERRCODE_OUT_OF_MEMORY, @@ -1916,43 +1916,22 @@ xml_parse(text *data, XmlOptionType xmloption_arg, "could not allocate XML document"); doc->standalone = standalone; - root = xmlNewNode(NULL, (const xmlChar *) "content-root"); - if (root == NULL || xmlerrcxt->err_occurred) - xml_ereport(xmlerrcxt, ERROR, ERRCODE_OUT_OF_MEMORY, - "could not allocate xml node"); - - /* - * This attaches root to doc, so we need not free it separately; - * and there can't yet be any old root to free. - */ - oldroot = xmlDocSetRootElement(doc, root); - Assert(oldroot == NULL); + /* set parse options --- have to do this the ugly way */ + save_keep_blanks = xmlKeepBlanksDefault(preserve_whitespace ? 1 : 0); /* allow empty content */ if (*(utf8string + count)) { - xmlNodePtr node_list = NULL; - xmlParserErrors res; - - res = xmlParseInNodeContext(root, - (char *) utf8string + count, - strlen((char *) utf8string + count), - options, - &node_list); - - if (res != XML_ERR_OK || xmlerrcxt->err_occurred) + res_code = xmlParseBalancedChunkMemory(doc, NULL, NULL, 0, + utf8string + count, + parsed_nodes); + if (res_code != 0 || xmlerrcxt->err_occurred) { - xmlFreeNodeList(node_list); xml_errsave(escontext, xmlerrcxt, ERRCODE_INVALID_XML_CONTENT, "invalid XML content"); goto fail; } - - if (parsed_nodes != NULL) - *parsed_nodes = node_list; - else - xmlFreeNodeList(node_list); } } @@ -1961,6 +1940,8 @@ xml_parse(text *data, XmlOptionType xmloption_arg, } PG_CATCH(); { + if (save_keep_blanks != -1) + xmlKeepBlanksDefault(save_keep_blanks); if (doc != NULL) xmlFreeDoc(doc); if (ctxt != NULL) @@ -1972,6 +1953,9 @@ xml_parse(text *data, XmlOptionType xmloption_arg, } PG_END_TRY(); + if (save_keep_blanks != -1) + xmlKeepBlanksDefault(save_keep_blanks); + if (ctxt != NULL) xmlFreeParserCtxt(ctxt); diff --git a/src/backend/utils/cache/evtcache.c b/src/backend/utils/cache/evtcache.c index ce596bf563856..b9d5a5998be50 100644 --- a/src/backend/utils/cache/evtcache.c +++ b/src/backend/utils/cache/evtcache.c @@ -78,7 +78,6 @@ BuildEventTriggerCache(void) { HASHCTL ctl; HTAB *cache; - MemoryContext oldcontext; Relation rel; Relation irel; SysScanDesc scan; @@ -110,9 +109,6 @@ BuildEventTriggerCache(void) (Datum) 0); } - /* Switch to correct memory context. */ - oldcontext = MemoryContextSwitchTo(EventTriggerCacheContext); - /* Prevent the memory context from being nuked while we're rebuilding. */ EventTriggerCacheState = ETCS_REBUILD_STARTED; @@ -145,6 +141,7 @@ BuildEventTriggerCache(void) bool evttags_isnull; EventTriggerCacheEntry *entry; bool found; + MemoryContext oldcontext; /* Get next tuple. */ tup = systable_getnext_ordered(scan, ForwardScanDirection); @@ -171,6 +168,9 @@ BuildEventTriggerCache(void) else continue; + /* Switch to correct memory context. */ + oldcontext = MemoryContextSwitchTo(EventTriggerCacheContext); + /* Allocate new cache item. */ item = palloc0(sizeof(EventTriggerCacheItem)); item->fnoid = form->evtfoid; @@ -188,6 +188,9 @@ BuildEventTriggerCache(void) entry->triggerlist = lappend(entry->triggerlist, item); else entry->triggerlist = list_make1(item); + + /* Restore previous memory context. */ + MemoryContextSwitchTo(oldcontext); } /* Done with pg_event_trigger scan. */ @@ -195,9 +198,6 @@ BuildEventTriggerCache(void) index_close(irel, AccessShareLock); relation_close(rel, AccessShareLock); - /* Restore previous memory context. */ - MemoryContextSwitchTo(oldcontext); - /* Install new cache. */ EventTriggerCache = cache; @@ -240,6 +240,8 @@ DecodeTextArrayToBitmapset(Datum array) } pfree(elems); + if ((Pointer) arr != DatumGetPointer(array)) + pfree(arr); return bms; } diff --git a/src/backend/utils/cache/plancache.c b/src/backend/utils/cache/plancache.c index 89a1c79e984d1..6661d2c6b7391 100644 --- a/src/backend/utils/cache/plancache.c +++ b/src/backend/utils/cache/plancache.c @@ -463,8 +463,7 @@ CompleteCachedPlan(CachedPlanSource *plansource, /* * Save the final parameter types (or other parameter specification data) - * into the source_context, as well as our other parameters. Also save - * the result tuple descriptor. + * into the source_context, as well as our other parameters. */ MemoryContextSwitchTo(source_context); @@ -480,9 +479,25 @@ CompleteCachedPlan(CachedPlanSource *plansource, plansource->parserSetupArg = parserSetupArg; plansource->cursor_options = cursor_options; plansource->fixed_result = fixed_result; - plansource->resultDesc = PlanCacheComputeResultDesc(querytree_list); + /* + * Also save the result tuple descriptor. PlanCacheComputeResultDesc may + * leak some cruft; normally we just accept that to save a copy step, but + * in USE_VALGRIND mode be tidy by running it in the caller's context. + */ +#ifdef USE_VALGRIND + MemoryContextSwitchTo(oldcxt); + plansource->resultDesc = PlanCacheComputeResultDesc(querytree_list); + if (plansource->resultDesc) + { + MemoryContextSwitchTo(source_context); + plansource->resultDesc = CreateTupleDescCopy(plansource->resultDesc); + MemoryContextSwitchTo(oldcxt); + } +#else + plansource->resultDesc = PlanCacheComputeResultDesc(querytree_list); MemoryContextSwitchTo(oldcxt); +#endif plansource->is_complete = true; plansource->is_valid = true; @@ -1283,6 +1298,7 @@ GetCachedPlan(CachedPlanSource *plansource, ParamListInfo boundParams, CachedPlan *plan = NULL; List *qlist; bool customplan; + ListCell *lc; /* Assert caller is doing things in a sane order */ Assert(plansource->magic == CACHEDPLANSOURCE_MAGIC); @@ -1385,6 +1401,13 @@ GetCachedPlan(CachedPlanSource *plansource, ParamListInfo boundParams, plan->is_saved = true; } + foreach(lc, plan->stmt_list) + { + PlannedStmt *pstmt = (PlannedStmt *) lfirst(lc); + + pstmt->planOrigin = customplan ? PLAN_STMT_CACHE_CUSTOM : PLAN_STMT_CACHE_GENERIC; + } + return plan; } diff --git a/src/backend/utils/cache/ts_cache.c b/src/backend/utils/cache/ts_cache.c index 18cccd778fd8c..e8ae53238d07a 100644 --- a/src/backend/utils/cache/ts_cache.c +++ b/src/backend/utils/cache/ts_cache.c @@ -321,7 +321,9 @@ lookup_ts_dictionary_cache(Oid dictId) /* * Init method runs in dictionary's private memory context, and we - * make sure the options are stored there too + * make sure the options are stored there too. This typically + * results in a small amount of memory leakage, but it's not worth + * complicating the API for tmplinit functions to avoid it. */ oldcontext = MemoryContextSwitchTo(entry->dictCtx); diff --git a/src/backend/utils/cache/typcache.c b/src/backend/utils/cache/typcache.c index f9aec38a11fb3..6a347698edffe 100644 --- a/src/backend/utils/cache/typcache.c +++ b/src/backend/utils/cache/typcache.c @@ -1171,9 +1171,6 @@ load_domaintype_info(TypeCacheEntry *typentry) elog(ERROR, "domain \"%s\" constraint \"%s\" has NULL conbin", NameStr(typTup->typname), NameStr(c->conname)); - /* Convert conbin to C string in caller context */ - constring = TextDatumGetCString(val); - /* Create the DomainConstraintCache object and context if needed */ if (dcc == NULL) { @@ -1189,9 +1186,8 @@ load_domaintype_info(TypeCacheEntry *typentry) dcc->dccRefCount = 0; } - /* Create node trees in DomainConstraintCache's context */ - oldcxt = MemoryContextSwitchTo(dcc->dccContext); - + /* Convert conbin to a node tree, still in caller's context */ + constring = TextDatumGetCString(val); check_expr = (Expr *) stringToNode(constring); /* @@ -1206,10 +1202,13 @@ load_domaintype_info(TypeCacheEntry *typentry) */ check_expr = expression_planner(check_expr); + /* Create only the minimally needed stuff in dccContext */ + oldcxt = MemoryContextSwitchTo(dcc->dccContext); + r = makeNode(DomainConstraintState); r->constrainttype = DOM_CONSTRAINT_CHECK; r->name = pstrdup(NameStr(c->conname)); - r->check_expr = check_expr; + r->check_expr = copyObject(check_expr); r->check_exprstate = NULL; MemoryContextSwitchTo(oldcxt); diff --git a/src/backend/utils/error/elog.c b/src/backend/utils/error/elog.c index 47af743990fe9..afce1a8e1f003 100644 --- a/src/backend/utils/error/elog.c +++ b/src/backend/utils/error/elog.c @@ -1128,12 +1128,15 @@ set_backtrace(ErrorData *edata, int num_skip) nframes = backtrace(buf, lengthof(buf)); strfrms = backtrace_symbols(buf, nframes); - if (strfrms == NULL) - return; - - for (int i = num_skip; i < nframes; i++) - appendStringInfo(&errtrace, "\n%s", strfrms[i]); - free(strfrms); + if (strfrms != NULL) + { + for (int i = num_skip; i < nframes; i++) + appendStringInfo(&errtrace, "\n%s", strfrms[i]); + free(strfrms); + } + else + appendStringInfoString(&errtrace, + "insufficient memory for backtrace generation"); } #else appendStringInfoString(&errtrace, diff --git a/src/backend/utils/hash/dynahash.c b/src/backend/utils/hash/dynahash.c index 1ad155d446e51..81da03629f0d2 100644 --- a/src/backend/utils/hash/dynahash.c +++ b/src/backend/utils/hash/dynahash.c @@ -22,10 +22,11 @@ * lookup key's hash value as a partition number --- this will work because * of the way calc_bucket() maps hash values to bucket numbers. * - * For hash tables in shared memory, the memory allocator function should - * match malloc's semantics of returning NULL on failure. For hash tables - * in local memory, we typically use palloc() which will throw error on - * failure. The code in this file has to cope with both cases. + * The memory allocator function should match malloc's semantics of returning + * NULL on failure. (This is essential for hash tables in shared memory. + * For hash tables in local memory, we used to use palloc() which will throw + * error on failure; but we no longer do, so it's untested whether this + * module will still cope with that behavior.) * * dynahash.c provides support for these types of lookup keys: * @@ -98,6 +99,7 @@ #include "access/xact.h" #include "common/hashfn.h" +#include "lib/ilist.h" #include "port/pg_bitutils.h" #include "storage/shmem.h" #include "storage/spin.h" @@ -195,6 +197,7 @@ struct HASHHDR long ssize; /* segment size --- must be power of 2 */ int sshift; /* segment shift = log2(ssize) */ int nelem_alloc; /* number of entries to allocate at once */ + bool isfixed; /* if true, don't enlarge */ #ifdef HASH_STATISTICS @@ -227,7 +230,6 @@ struct HTAB MemoryContext hcxt; /* memory context if default allocator used */ char *tabname; /* table name (for error messages) */ bool isshared; /* true if table is in shared memory */ - bool isfixed; /* if true, don't enlarge */ /* freezing a shared table isn't allowed, so we can keep state here */ bool frozen; /* true = no more inserts allowed */ @@ -236,6 +238,16 @@ struct HTAB Size keysize; /* hash key length in bytes */ long ssize; /* segment size --- must be power of 2 */ int sshift; /* segment shift = log2(ssize) */ + + /* + * In a USE_VALGRIND build, non-shared hashtables keep an slist chain of + * all the element blocks they have allocated. This pacifies Valgrind, + * which would otherwise often claim that the element blocks are "possibly + * lost" for lack of any non-interior pointers to their starts. + */ +#ifdef USE_VALGRIND + slist_head element_blocks; +#endif }; /* @@ -618,8 +630,10 @@ hash_create(const char *tabname, long nelem, const HASHCTL *info, int flags) } } + /* Set isfixed if requested, but not till after we build initial entries */ if (flags & HASH_FIXED_SIZE) - hashp->isfixed = true; + hctl->isfixed = true; + return hashp; } @@ -644,6 +658,8 @@ hdefault(HTAB *hashp) hctl->ssize = DEF_SEGSIZE; hctl->sshift = DEF_SEGSIZE_SHIFT; + hctl->isfixed = false; /* can be enlarged */ + #ifdef HASH_STATISTICS hctl->accesses = hctl->collisions = 0; #endif @@ -1708,23 +1724,51 @@ element_alloc(HTAB *hashp, int nelem, int freelist_idx) { HASHHDR *hctl = hashp->hctl; Size elementSize; + Size requestSize; + char *allocedBlock; HASHELEMENT *firstElement; HASHELEMENT *tmpElement; HASHELEMENT *prevElement; int i; - if (hashp->isfixed) + if (hctl->isfixed) return false; /* Each element has a HASHELEMENT header plus user data. */ elementSize = MAXALIGN(sizeof(HASHELEMENT)) + MAXALIGN(hctl->entrysize); + requestSize = nelem * elementSize; + + /* Add space for slist_node list link if we need one. */ +#ifdef USE_VALGRIND + if (!hashp->isshared) + requestSize += MAXALIGN(sizeof(slist_node)); +#endif + + /* Allocate the memory. */ CurrentDynaHashCxt = hashp->hcxt; - firstElement = (HASHELEMENT *) hashp->alloc(nelem * elementSize); + allocedBlock = hashp->alloc(requestSize); - if (!firstElement) + if (!allocedBlock) return false; + /* + * If USE_VALGRIND, each allocated block of elements of a non-shared + * hashtable is chained into a list, so that Valgrind won't think it's + * been leaked. + */ +#ifdef USE_VALGRIND + if (hashp->isshared) + firstElement = (HASHELEMENT *) allocedBlock; + else + { + slist_push_head(&hashp->element_blocks, (slist_node *) allocedBlock); + firstElement = (HASHELEMENT *) (allocedBlock + MAXALIGN(sizeof(slist_node))); + } +#else + firstElement = (HASHELEMENT *) allocedBlock; +#endif + /* prepare to link all the new entries into the freelist */ prevElement = NULL; tmpElement = firstElement; diff --git a/src/backend/utils/init/miscinit.c b/src/backend/utils/init/miscinit.c index 43b4dbccc3de6..65d8cbfaed585 100644 --- a/src/backend/utils/init/miscinit.c +++ b/src/backend/utils/init/miscinit.c @@ -1183,7 +1183,6 @@ UnlinkLockFiles(int status, Datum arg) /* Should we complain if the unlink fails? */ } /* Since we're about to exit, no need to reclaim storage */ - lock_files = NIL; /* * Lock file removal should always be the last externally visible action diff --git a/src/backend/utils/misc/guc.c b/src/backend/utils/misc/guc.c index 667df448732f2..e404c345e6eab 100644 --- a/src/backend/utils/misc/guc.c +++ b/src/backend/utils/misc/guc.c @@ -249,6 +249,7 @@ static void reapply_stacked_values(struct config_generic *variable, const char *curvalue, GucContext curscontext, GucSource cursource, Oid cursrole); +static void free_placeholder(struct config_string *pHolder); static bool validate_option_array_item(const char *name, const char *value, bool skipIfNoPermissions); static void write_auto_conf_file(int fd, const char *filename, ConfigVariable *head); @@ -4722,8 +4723,13 @@ AlterSystemSetConfigFile(AlterSystemStmt *altersysstmt) * the config file cannot cause postmaster start to fail, so we * don't have to be too tense about possibly installing a bad * value.) + * + * As an exception, we skip this check if this is a RESET command + * for an unknown custom GUC, else there'd be no way for users to + * remove such settings with reserved prefixes. */ - (void) assignable_custom_variable_name(name, false, ERROR); + if (value || !valid_custom_variable_name(name)) + (void) assignable_custom_variable_name(name, false, ERROR); } /* @@ -5018,16 +5024,8 @@ define_custom_variable(struct config_generic *variable) set_config_sourcefile(name, pHolder->gen.sourcefile, pHolder->gen.sourceline); - /* - * Free up as much as we conveniently can of the placeholder structure. - * (This neglects any stack items, so it's possible for some memory to be - * leaked. Since this can only happen once per session per variable, it - * doesn't seem worth spending much code on.) - */ - set_string_field(pHolder, pHolder->variable, NULL); - set_string_field(pHolder, &pHolder->reset_val, NULL); - - guc_free(pHolder); + /* Now we can free the no-longer-referenced placeholder variable */ + free_placeholder(pHolder); } /* @@ -5126,6 +5124,25 @@ reapply_stacked_values(struct config_generic *variable, } } +/* + * Free up a no-longer-referenced placeholder GUC variable. + * + * This neglects any stack items, so it's possible for some memory to be + * leaked. Since this can only happen once per session per variable, it + * doesn't seem worth spending much code on. + */ +static void +free_placeholder(struct config_string *pHolder) +{ + /* Placeholders are always STRING type, so free their values */ + Assert(pHolder->gen.vartype == PGC_STRING); + set_string_field(pHolder, pHolder->variable, NULL); + set_string_field(pHolder, &pHolder->reset_val, NULL); + + guc_free(unconstify(char *, pHolder->gen.name)); + guc_free(pHolder); +} + /* * Functions for extensions to call to define their custom GUC variables. */ @@ -5286,9 +5303,7 @@ MarkGUCPrefixReserved(const char *className) /* * Check for existing placeholders. We must actually remove invalid - * placeholders, else future parallel worker startups will fail. (We - * don't bother trying to free associated memory, since this shouldn't - * happen often.) + * placeholders, else future parallel worker startups will fail. */ hash_seq_init(&status, guc_hashtab); while ((hentry = (GUCHashEntry *) hash_seq_search(&status)) != NULL) @@ -5312,6 +5327,8 @@ MarkGUCPrefixReserved(const char *className) NULL); /* Remove it from any lists it's in, too */ RemoveGUCFromLists(var); + /* And free it */ + free_placeholder((struct config_string *) var); } } @@ -6711,6 +6728,7 @@ validate_option_array_item(const char *name, const char *value, { struct config_generic *gconf; + bool reset_custom; /* * There are three cases to consider: @@ -6729,16 +6747,21 @@ validate_option_array_item(const char *name, const char *value, * it's assumed to be fully validated.) * * name is not known and can't be created as a placeholder. Throw error, - * unless skipIfNoPermissions is true, in which case return false. + * unless skipIfNoPermissions or reset_custom is true. If reset_custom is + * true, this is a RESET or RESET ALL operation for an unknown custom GUC + * with a reserved prefix, in which case we want to fall through to the + * placeholder case described in the preceding paragraph (else there'd be + * no way for users to remove them). Otherwise, return false. */ - gconf = find_option(name, true, skipIfNoPermissions, ERROR); - if (!gconf) + reset_custom = (!value && valid_custom_variable_name(name)); + gconf = find_option(name, true, skipIfNoPermissions || reset_custom, ERROR); + if (!gconf && !reset_custom) { /* not known, failed to make a placeholder */ return false; } - if (gconf->flags & GUC_CUSTOM_PLACEHOLDER) + if (!gconf || gconf->flags & GUC_CUSTOM_PLACEHOLDER) { /* * We cannot do any meaningful check on the value, so only permissions diff --git a/src/backend/utils/misc/ps_status.c b/src/backend/utils/misc/ps_status.c index e08b26e8c14f2..4df25944deb33 100644 --- a/src/backend/utils/misc/ps_status.c +++ b/src/backend/utils/misc/ps_status.c @@ -100,6 +100,17 @@ static void flush_ps_display(void); static int save_argc; static char **save_argv; +/* + * Valgrind seems not to consider the global "environ" variable as a valid + * root pointer; so when we allocate a new environment array, it claims that + * data is leaked. To fix that, keep our own statically-allocated copy of the + * pointer. (Oddly, this doesn't seem to be a problem for "argv".) + */ +#if defined(PS_USE_CLOBBER_ARGV) && defined(USE_VALGRIND) +extern char **ps_status_new_environ; +char **ps_status_new_environ; +#endif + /* * Call this early in startup to save the original argc/argv values. @@ -206,6 +217,11 @@ save_ps_display_args(int argc, char **argv) } new_environ[i] = NULL; environ = new_environ; + + /* See notes about Valgrind above. */ +#ifdef USE_VALGRIND + ps_status_new_environ = new_environ; +#endif } /* diff --git a/src/backend/utils/mmgr/alignedalloc.c b/src/backend/utils/mmgr/alignedalloc.c index 7eea695de62c5..b1be742691497 100644 --- a/src/backend/utils/mmgr/alignedalloc.c +++ b/src/backend/utils/mmgr/alignedalloc.c @@ -45,6 +45,15 @@ AlignedAllocFree(void *pointer) GetMemoryChunkContext(unaligned)->name, chunk); #endif + /* + * Create a dummy vchunk covering the start of the unaligned chunk, but + * not overlapping the aligned chunk. This will be freed while pfree'ing + * the unaligned chunk, keeping Valgrind happy. Then when we return to + * the outer pfree, that will clean up the vchunk for the aligned chunk. + */ + VALGRIND_MEMPOOL_ALLOC(GetMemoryChunkContext(unaligned), unaligned, + (char *) pointer - (char *) unaligned); + /* Recursively pfree the unaligned chunk */ pfree(unaligned); } @@ -123,6 +132,15 @@ AlignedAllocRealloc(void *pointer, Size size, int flags) VALGRIND_MAKE_MEM_DEFINED(pointer, old_size); memcpy(newptr, pointer, Min(size, old_size)); + /* + * Create a dummy vchunk covering the start of the old unaligned chunk, + * but not overlapping the aligned chunk. This will be freed while + * pfree'ing the old unaligned chunk, keeping Valgrind happy. Then when + * we return to repalloc, it will move the vchunk for the aligned chunk. + */ + VALGRIND_MEMPOOL_ALLOC(ctx, unaligned, + (char *) pointer - (char *) unaligned); + pfree(unaligned); return newptr; diff --git a/src/backend/utils/mmgr/aset.c b/src/backend/utils/mmgr/aset.c index 666ecd8f78d0e..9ef109ca586bd 100644 --- a/src/backend/utils/mmgr/aset.c +++ b/src/backend/utils/mmgr/aset.c @@ -103,6 +103,8 @@ #define ALLOC_BLOCKHDRSZ MAXALIGN(sizeof(AllocBlockData)) #define ALLOC_CHUNKHDRSZ sizeof(MemoryChunk) +#define FIRST_BLOCKHDRSZ (MAXALIGN(sizeof(AllocSetContext)) + \ + ALLOC_BLOCKHDRSZ) typedef struct AllocBlockData *AllocBlock; /* forward reference */ @@ -458,6 +460,21 @@ AllocSetContextCreateInternal(MemoryContext parent, * we'd leak the header/initial block if we ereport in this stretch. */ + /* Create a vpool associated with the context */ + VALGRIND_CREATE_MEMPOOL(set, 0, false); + + /* + * Create a vchunk covering both the AllocSetContext struct and the keeper + * block's header. (Perhaps it would be more sensible for these to be two + * separate vchunks, but doing that seems to tickle bugs in some versions + * of Valgrind.) We must have these vchunks, and also a vchunk for each + * subsequently-added block header, so that Valgrind considers the + * pointers within them while checking for leaked memory. Note that + * Valgrind doesn't distinguish between these vchunks and those created by + * mcxt.c for the user-accessible-data chunks we allocate. + */ + VALGRIND_MEMPOOL_ALLOC(set, set, FIRST_BLOCKHDRSZ); + /* Fill in the initial block's block header */ block = KeeperBlock(set); block->aset = set; @@ -585,6 +602,14 @@ AllocSetReset(MemoryContext context) #ifdef CLOBBER_FREED_MEMORY wipe_mem(block, block->freeptr - ((char *) block)); #endif + + /* + * We need to free the block header's vchunk explicitly, although + * the user-data vchunks within will go away in the TRIM below. + * Otherwise Valgrind complains about leaked allocations. + */ + VALGRIND_MEMPOOL_FREE(set, block); + free(block); } block = next; @@ -592,6 +617,14 @@ AllocSetReset(MemoryContext context) Assert(context->mem_allocated == keepersize); + /* + * Instruct Valgrind to throw away all the vchunks associated with this + * context, except for the one covering the AllocSetContext and + * keeper-block header. This gets rid of the vchunks for whatever user + * data is getting discarded by the context reset. + */ + VALGRIND_MEMPOOL_TRIM(set, set, FIRST_BLOCKHDRSZ); + /* Reset block size allocation sequence, too */ set->nextBlockSize = set->initBlockSize; } @@ -648,6 +681,9 @@ AllocSetDelete(MemoryContext context) freelist->first_free = (AllocSetContext *) oldset->header.nextchild; freelist->num_free--; + /* Destroy the context's vpool --- see notes below */ + VALGRIND_DESTROY_MEMPOOL(oldset); + /* All that remains is to free the header/initial block */ free(oldset); } @@ -675,13 +711,24 @@ AllocSetDelete(MemoryContext context) #endif if (!IsKeeperBlock(set, block)) + { + /* As in AllocSetReset, free block-header vchunks explicitly */ + VALGRIND_MEMPOOL_FREE(set, block); free(block); + } block = next; } Assert(context->mem_allocated == keepersize); + /* + * Destroy the vpool. We don't seem to need to explicitly free the + * initial block's header vchunk, nor any user-data vchunks that Valgrind + * still knows about; they'll all go away automatically. + */ + VALGRIND_DESTROY_MEMPOOL(set); + /* Finally, free the context header, including the keeper block */ free(set); } @@ -716,6 +763,9 @@ AllocSetAllocLarge(MemoryContext context, Size size, int flags) if (block == NULL) return MemoryContextAllocationFailure(context, size, flags); + /* Make a vchunk covering the new block's header */ + VALGRIND_MEMPOOL_ALLOC(set, block, ALLOC_BLOCKHDRSZ); + context->mem_allocated += blksize; block->aset = set; @@ -922,6 +972,9 @@ AllocSetAllocFromNewBlock(MemoryContext context, Size size, int flags, if (block == NULL) return MemoryContextAllocationFailure(context, size, flags); + /* Make a vchunk covering the new block's header */ + VALGRIND_MEMPOOL_ALLOC(set, block, ALLOC_BLOCKHDRSZ); + context->mem_allocated += blksize; block->aset = set; @@ -1104,6 +1157,10 @@ AllocSetFree(void *pointer) #ifdef CLOBBER_FREED_MEMORY wipe_mem(block, block->freeptr - ((char *) block)); #endif + + /* As in AllocSetReset, free block-header vchunks explicitly */ + VALGRIND_MEMPOOL_FREE(set, block); + free(block); } else @@ -1184,6 +1241,7 @@ AllocSetRealloc(void *pointer, Size size, int flags) * realloc() to make the containing block bigger, or smaller, with * minimum space wastage. */ + AllocBlock newblock; Size chksize; Size blksize; Size oldblksize; @@ -1223,14 +1281,21 @@ AllocSetRealloc(void *pointer, Size size, int flags) blksize = chksize + ALLOC_BLOCKHDRSZ + ALLOC_CHUNKHDRSZ; oldblksize = block->endptr - ((char *) block); - block = (AllocBlock) realloc(block, blksize); - if (block == NULL) + newblock = (AllocBlock) realloc(block, blksize); + if (newblock == NULL) { /* Disallow access to the chunk header. */ VALGRIND_MAKE_MEM_NOACCESS(chunk, ALLOC_CHUNKHDRSZ); return MemoryContextAllocationFailure(&set->header, size, flags); } + /* + * Move the block-header vchunk explicitly. (mcxt.c will take care of + * moving the vchunk for the user data.) + */ + VALGRIND_MEMPOOL_CHANGE(set, block, newblock, ALLOC_BLOCKHDRSZ); + block = newblock; + /* updated separately, not to underflow when (oldblksize > blksize) */ set->header.mem_allocated -= oldblksize; set->header.mem_allocated += blksize; @@ -1294,7 +1359,7 @@ AllocSetRealloc(void *pointer, Size size, int flags) /* Ensure any padding bytes are marked NOACCESS. */ VALGRIND_MAKE_MEM_NOACCESS((char *) pointer + size, chksize - size); - /* Disallow access to the chunk header . */ + /* Disallow access to the chunk header. */ VALGRIND_MAKE_MEM_NOACCESS(chunk, ALLOC_CHUNKHDRSZ); return pointer; diff --git a/src/backend/utils/mmgr/bump.c b/src/backend/utils/mmgr/bump.c index f7a37d1b3e86c..2805d55a2eca9 100644 --- a/src/backend/utils/mmgr/bump.c +++ b/src/backend/utils/mmgr/bump.c @@ -45,7 +45,9 @@ #include "utils/memutils_memorychunk.h" #include "utils/memutils_internal.h" -#define Bump_BLOCKHDRSZ MAXALIGN(sizeof(BumpBlock)) +#define Bump_BLOCKHDRSZ MAXALIGN(sizeof(BumpBlock)) +#define FIRST_BLOCKHDRSZ (MAXALIGN(sizeof(BumpContext)) + \ + Bump_BLOCKHDRSZ) /* No chunk header unless built with MEMORY_CONTEXT_CHECKING */ #ifdef MEMORY_CONTEXT_CHECKING @@ -189,6 +191,12 @@ BumpContextCreate(MemoryContext parent, const char *name, Size minContextSize, * Avoid writing code that can fail between here and MemoryContextCreate; * we'd leak the header and initial block if we ereport in this stretch. */ + + /* See comments about Valgrind interactions in aset.c */ + VALGRIND_CREATE_MEMPOOL(set, 0, false); + /* This vchunk covers the BumpContext and the keeper block header */ + VALGRIND_MEMPOOL_ALLOC(set, set, FIRST_BLOCKHDRSZ); + dlist_init(&set->blocks); /* Fill in the initial block's block header */ @@ -262,6 +270,14 @@ BumpReset(MemoryContext context) BumpBlockFree(set, block); } + /* + * Instruct Valgrind to throw away all the vchunks associated with this + * context, except for the one covering the BumpContext and keeper-block + * header. This gets rid of the vchunks for whatever user data is getting + * discarded by the context reset. + */ + VALGRIND_MEMPOOL_TRIM(set, set, FIRST_BLOCKHDRSZ); + /* Reset block size allocation sequence, too */ set->nextBlockSize = set->initBlockSize; @@ -279,6 +295,10 @@ BumpDelete(MemoryContext context) { /* Reset to release all releasable BumpBlocks */ BumpReset(context); + + /* Destroy the vpool -- see notes in aset.c */ + VALGRIND_DESTROY_MEMPOOL(context); + /* And free the context header and keeper block */ free(context); } @@ -318,6 +338,9 @@ BumpAllocLarge(MemoryContext context, Size size, int flags) if (block == NULL) return MemoryContextAllocationFailure(context, size, flags); + /* Make a vchunk covering the new block's header */ + VALGRIND_MEMPOOL_ALLOC(set, block, Bump_BLOCKHDRSZ); + context->mem_allocated += blksize; /* the block is completely full */ @@ -455,6 +478,9 @@ BumpAllocFromNewBlock(MemoryContext context, Size size, int flags, if (block == NULL) return MemoryContextAllocationFailure(context, size, flags); + /* Make a vchunk covering the new block's header */ + VALGRIND_MEMPOOL_ALLOC(set, block, Bump_BLOCKHDRSZ); + context->mem_allocated += blksize; /* initialize the new block */ @@ -606,6 +632,9 @@ BumpBlockFree(BumpContext *set, BumpBlock *block) wipe_mem(block, ((char *) block->endptr - (char *) block)); #endif + /* As in aset.c, free block-header vchunks explicitly */ + VALGRIND_MEMPOOL_FREE(set, block); + free(block); } diff --git a/src/backend/utils/mmgr/generation.c b/src/backend/utils/mmgr/generation.c index 18679ad4f1e41..cfafc9bf0829d 100644 --- a/src/backend/utils/mmgr/generation.c +++ b/src/backend/utils/mmgr/generation.c @@ -45,6 +45,8 @@ #define Generation_BLOCKHDRSZ MAXALIGN(sizeof(GenerationBlock)) #define Generation_CHUNKHDRSZ sizeof(MemoryChunk) +#define FIRST_BLOCKHDRSZ (MAXALIGN(sizeof(GenerationContext)) + \ + Generation_BLOCKHDRSZ) #define Generation_CHUNK_FRACTION 8 @@ -221,6 +223,12 @@ GenerationContextCreate(MemoryContext parent, * Avoid writing code that can fail between here and MemoryContextCreate; * we'd leak the header if we ereport in this stretch. */ + + /* See comments about Valgrind interactions in aset.c */ + VALGRIND_CREATE_MEMPOOL(set, 0, false); + /* This vchunk covers the GenerationContext and the keeper block header */ + VALGRIND_MEMPOOL_ALLOC(set, set, FIRST_BLOCKHDRSZ); + dlist_init(&set->blocks); /* Fill in the initial block's block header */ @@ -309,6 +317,14 @@ GenerationReset(MemoryContext context) GenerationBlockFree(set, block); } + /* + * Instruct Valgrind to throw away all the vchunks associated with this + * context, except for the one covering the GenerationContext and + * keeper-block header. This gets rid of the vchunks for whatever user + * data is getting discarded by the context reset. + */ + VALGRIND_MEMPOOL_TRIM(set, set, FIRST_BLOCKHDRSZ); + /* set it so new allocations to make use of the keeper block */ set->block = KeeperBlock(set); @@ -329,6 +345,10 @@ GenerationDelete(MemoryContext context) { /* Reset to release all releasable GenerationBlocks */ GenerationReset(context); + + /* Destroy the vpool -- see notes in aset.c */ + VALGRIND_DESTROY_MEMPOOL(context); + /* And free the context header and keeper block */ free(context); } @@ -365,6 +385,9 @@ GenerationAllocLarge(MemoryContext context, Size size, int flags) if (block == NULL) return MemoryContextAllocationFailure(context, size, flags); + /* Make a vchunk covering the new block's header */ + VALGRIND_MEMPOOL_ALLOC(set, block, Generation_BLOCKHDRSZ); + context->mem_allocated += blksize; /* block with a single (used) chunk */ @@ -487,6 +510,9 @@ GenerationAllocFromNewBlock(MemoryContext context, Size size, int flags, if (block == NULL) return MemoryContextAllocationFailure(context, size, flags); + /* Make a vchunk covering the new block's header */ + VALGRIND_MEMPOOL_ALLOC(set, block, Generation_BLOCKHDRSZ); + context->mem_allocated += blksize; /* initialize the new block */ @@ -677,6 +703,9 @@ GenerationBlockFree(GenerationContext *set, GenerationBlock *block) wipe_mem(block, block->blksize); #endif + /* As in aset.c, free block-header vchunks explicitly */ + VALGRIND_MEMPOOL_FREE(set, block); + free(block); } diff --git a/src/backend/utils/mmgr/mcxt.c b/src/backend/utils/mmgr/mcxt.c index 15fa4d0a55eeb..47fd774c7d280 100644 --- a/src/backend/utils/mmgr/mcxt.c +++ b/src/backend/utils/mmgr/mcxt.c @@ -8,6 +8,23 @@ * context-type-specific operations via the function pointers in a * context's MemoryContextMethods struct. * + * A note about Valgrind support: when USE_VALGRIND is defined, we provide + * support for memory leak tracking at the allocation-unit level. Valgrind + * does leak detection by tracking allocated "chunks", which can be grouped + * into "pools". The "chunk" terminology is overloaded, since we use that + * word for our allocation units, and it's sometimes important to distinguish + * those from the Valgrind objects that describe them. To reduce confusion, + * let's use the terms "vchunk" and "vpool" for the Valgrind objects. + * + * We use a separate vpool for each memory context. The context-type-specific + * code is responsible for creating and deleting the vpools, and also for + * creating vchunks to cover its management data structures such as block + * headers. (There must be a vchunk that includes every pointer we want + * Valgrind to consider for leak-tracking purposes.) This module creates + * and deletes the vchunks that cover the caller-visible allocated chunks. + * However, the context-type-specific code must handle cleaning up those + * vchunks too during memory context reset operations. + * * * Portions Copyright (c) 1996-2025, PostgreSQL Global Development Group * Portions Copyright (c) 1994, Regents of the University of California @@ -418,8 +435,6 @@ MemoryContextResetOnly(MemoryContext context) context->methods->reset(context); context->isReset = true; - VALGRIND_DESTROY_MEMPOOL(context); - VALGRIND_CREATE_MEMPOOL(context, 0, false); } } @@ -526,8 +541,6 @@ MemoryContextDeleteOnly(MemoryContext context) context->ident = NULL; context->methods->delete_context(context); - - VALGRIND_DESTROY_MEMPOOL(context); } /* @@ -560,9 +573,7 @@ MemoryContextDeleteChildren(MemoryContext context) * the specified context, since that means it will automatically be freed * when no longer needed. * - * There is no API for deregistering a callback once registered. If you - * want it to not do anything anymore, adjust the state pointed to by its - * "arg" to indicate that. + * Note that callers can assume this cannot fail. */ void MemoryContextRegisterResetCallback(MemoryContext context, @@ -577,6 +588,41 @@ MemoryContextRegisterResetCallback(MemoryContext context, context->isReset = false; } +/* + * MemoryContextUnregisterResetCallback + * Undo the effects of MemoryContextRegisterResetCallback. + * + * This can be used if a callback's effects are no longer required + * at some point before the context has been reset/deleted. It is the + * caller's responsibility to pfree the callback struct (if needed). + * + * An assertion failure occurs if the callback was not registered. + * We could alternatively define that case as a no-op, but that seems too + * likely to mask programming errors such as passing the wrong context. + */ +void +MemoryContextUnregisterResetCallback(MemoryContext context, + MemoryContextCallback *cb) +{ + MemoryContextCallback *prev, + *cur; + + Assert(MemoryContextIsValid(context)); + + for (prev = NULL, cur = context->reset_cbs; cur != NULL; + prev = cur, cur = cur->next) + { + if (cur != cb) + continue; + if (prev) + prev->next = cur->next; + else + context->reset_cbs = cur->next; + return; + } + Assert(false); +} + /* * MemoryContextCallResetCallbacks * Internal function to call all registered callbacks for context. @@ -1137,8 +1183,6 @@ MemoryContextCreate(MemoryContext node, node->nextchild = NULL; node->allowInCritSection = false; } - - VALGRIND_CREATE_MEMPOOL(node, 0, false); } /* @@ -1421,7 +1465,13 @@ MemoryContextAllocAligned(MemoryContext context, void *unaligned; void *aligned; - /* wouldn't make much sense to waste that much space */ + /* + * Restrict alignto to ensure that it can fit into the "value" field of + * the redirection MemoryChunk, and that the distance back to the start of + * the unaligned chunk will fit into the space available for that. This + * isn't a limitation in practice, since it wouldn't make much sense to + * waste that much space. + */ Assert(alignto < (128 * 1024 * 1024)); /* ensure alignto is a power of 2 */ @@ -1458,10 +1508,15 @@ MemoryContextAllocAligned(MemoryContext context, alloc_size += 1; #endif - /* perform the actual allocation */ - unaligned = MemoryContextAllocExtended(context, alloc_size, flags); + /* + * Perform the actual allocation, but do not pass down MCXT_ALLOC_ZERO. + * This ensures that wasted bytes beyond the aligned chunk do not become + * DEFINED. + */ + unaligned = MemoryContextAllocExtended(context, alloc_size, + flags & ~MCXT_ALLOC_ZERO); - /* set the aligned pointer */ + /* compute the aligned pointer */ aligned = (void *) TYPEALIGN(alignto, (char *) unaligned + sizeof(MemoryChunk)); @@ -1489,12 +1544,23 @@ MemoryContextAllocAligned(MemoryContext context, set_sentinel(aligned, size); #endif - /* Mark the bytes before the redirection header as noaccess */ - VALGRIND_MAKE_MEM_NOACCESS(unaligned, - (char *) alignedchunk - (char *) unaligned); + /* + * MemoryContextAllocExtended marked the whole unaligned chunk as a + * vchunk. Undo that, instead making just the aligned chunk be a vchunk. + * This prevents Valgrind from complaining that the vchunk is possibly + * leaked, since only pointers to the aligned chunk will exist. + * + * After these calls, the aligned chunk will be marked UNDEFINED, and all + * the rest of the unaligned chunk (the redirection chunk header, the + * padding bytes before it, and any wasted trailing bytes) will be marked + * NOACCESS, which is what we want. + */ + VALGRIND_MEMPOOL_FREE(context, unaligned); + VALGRIND_MEMPOOL_ALLOC(context, aligned, size); - /* Disallow access to the redirection chunk header. */ - VALGRIND_MAKE_MEM_NOACCESS(alignedchunk, sizeof(MemoryChunk)); + /* Now zero (and make DEFINED) just the aligned chunk, if requested */ + if ((flags & MCXT_ALLOC_ZERO) != 0) + MemSetAligned(aligned, 0, size); return aligned; } @@ -1528,16 +1594,12 @@ void pfree(void *pointer) { #ifdef USE_VALGRIND - MemoryContextMethodID method = GetMemoryChunkMethodID(pointer); MemoryContext context = GetMemoryChunkContext(pointer); #endif MCXT_METHOD(pointer, free_p) (pointer); -#ifdef USE_VALGRIND - if (method != MCTX_ALIGNED_REDIRECT_ID) - VALGRIND_MEMPOOL_FREE(context, pointer); -#endif + VALGRIND_MEMPOOL_FREE(context, pointer); } /* @@ -1547,9 +1609,6 @@ pfree(void *pointer) void * repalloc(void *pointer, Size size) { -#ifdef USE_VALGRIND - MemoryContextMethodID method = GetMemoryChunkMethodID(pointer); -#endif #if defined(USE_ASSERT_CHECKING) || defined(USE_VALGRIND) MemoryContext context = GetMemoryChunkContext(pointer); #endif @@ -1572,10 +1631,7 @@ repalloc(void *pointer, Size size) */ ret = MCXT_METHOD(pointer, realloc) (pointer, size, 0); -#ifdef USE_VALGRIND - if (method != MCTX_ALIGNED_REDIRECT_ID) - VALGRIND_MEMPOOL_CHANGE(context, pointer, ret, size); -#endif + VALGRIND_MEMPOOL_CHANGE(context, pointer, ret, size); return ret; } diff --git a/src/backend/utils/mmgr/slab.c b/src/backend/utils/mmgr/slab.c index d32c0d318fbf4..0e35abcf5a055 100644 --- a/src/backend/utils/mmgr/slab.c +++ b/src/backend/utils/mmgr/slab.c @@ -377,6 +377,11 @@ SlabContextCreate(MemoryContext parent, * we'd leak the header if we ereport in this stretch. */ + /* See comments about Valgrind interactions in aset.c */ + VALGRIND_CREATE_MEMPOOL(slab, 0, false); + /* This vchunk covers the SlabContext only */ + VALGRIND_MEMPOOL_ALLOC(slab, slab, sizeof(SlabContext)); + /* Fill in SlabContext-specific header fields */ slab->chunkSize = (uint32) chunkSize; slab->fullChunkSize = (uint32) fullChunkSize; @@ -451,6 +456,10 @@ SlabReset(MemoryContext context) #ifdef CLOBBER_FREED_MEMORY wipe_mem(block, slab->blockSize); #endif + + /* As in aset.c, free block-header vchunks explicitly */ + VALGRIND_MEMPOOL_FREE(slab, block); + free(block); context->mem_allocated -= slab->blockSize; } @@ -467,11 +476,23 @@ SlabReset(MemoryContext context) #ifdef CLOBBER_FREED_MEMORY wipe_mem(block, slab->blockSize); #endif + + /* As in aset.c, free block-header vchunks explicitly */ + VALGRIND_MEMPOOL_FREE(slab, block); + free(block); context->mem_allocated -= slab->blockSize; } } + /* + * Instruct Valgrind to throw away all the vchunks associated with this + * context, except for the one covering the SlabContext. This gets rid of + * the vchunks for whatever user data is getting discarded by the context + * reset. + */ + VALGRIND_MEMPOOL_TRIM(slab, slab, sizeof(SlabContext)); + slab->curBlocklistIndex = 0; Assert(context->mem_allocated == 0); @@ -486,6 +507,10 @@ SlabDelete(MemoryContext context) { /* Reset to release all the SlabBlocks */ SlabReset(context); + + /* Destroy the vpool -- see notes in aset.c */ + VALGRIND_DESTROY_MEMPOOL(context); + /* And free the context header */ free(context); } @@ -567,6 +592,9 @@ SlabAllocFromNewBlock(MemoryContext context, Size size, int flags) if (unlikely(block == NULL)) return MemoryContextAllocationFailure(context, size, flags); + /* Make a vchunk covering the new block's header */ + VALGRIND_MEMPOOL_ALLOC(slab, block, Slab_BLOCKHDRSZ); + block->slab = slab; context->mem_allocated += slab->blockSize; @@ -795,6 +823,10 @@ SlabFree(void *pointer) #ifdef CLOBBER_FREED_MEMORY wipe_mem(block, slab->blockSize); #endif + + /* As in aset.c, free block-header vchunks explicitly */ + VALGRIND_MEMPOOL_FREE(slab, block); + free(block); slab->header.mem_allocated -= slab->blockSize; } diff --git a/src/bin/initdb/Makefile b/src/bin/initdb/Makefile index 997e0a013e956..c0470efda92a3 100644 --- a/src/bin/initdb/Makefile +++ b/src/bin/initdb/Makefile @@ -20,7 +20,7 @@ include $(top_builddir)/src/Makefile.global # from libpq, else we have risks of version skew if we run with a libpq # shared library from a different PG version. Define # USE_PRIVATE_ENCODING_FUNCS to ensure that that happens. -override CPPFLAGS := -DUSE_PRIVATE_ENCODING_FUNCS -I$(libpq_srcdir) -I$(top_srcdir)/src/timezone $(ICU_CFLAGS) $(CPPFLAGS) +override CPPFLAGS := -DUSE_PRIVATE_ENCODING_FUNCS -I$(libpq_srcdir) -I$(top_srcdir)/src/timezone $(CPPFLAGS) $(ICU_CFLAGS) # We need libpq only because fe_utils does. LDFLAGS_INTERNAL += -L$(top_builddir)/src/fe_utils -lpgfeutils $(libpq_pgport) $(ICU_LIBS) diff --git a/src/bin/pg_basebackup/pg_createsubscriber.c b/src/bin/pg_basebackup/pg_createsubscriber.c index 025b893a41e83..3986882f04292 100644 --- a/src/bin/pg_basebackup/pg_createsubscriber.c +++ b/src/bin/pg_basebackup/pg_createsubscriber.c @@ -1250,8 +1250,17 @@ setup_recovery(const struct LogicalRepInfo *dbinfo, const char *datadir, const c appendPQExpBufferStr(recoveryconfcontents, "recovery_target = ''\n"); appendPQExpBufferStr(recoveryconfcontents, "recovery_target_timeline = 'latest'\n"); + + /* + * Set recovery_target_inclusive = false to avoid reapplying the + * transaction committed at 'lsn' after subscription is enabled. This is + * because the provided 'lsn' is also used as the replication start point + * for the subscription. So, the server can send the transaction committed + * at that 'lsn' after replication is started which can lead to applying + * the same transaction twice if we keep recovery_target_inclusive = true. + */ appendPQExpBufferStr(recoveryconfcontents, - "recovery_target_inclusive = true\n"); + "recovery_target_inclusive = false\n"); appendPQExpBufferStr(recoveryconfcontents, "recovery_target_action = promote\n"); appendPQExpBufferStr(recoveryconfcontents, "recovery_target_name = ''\n"); diff --git a/src/bin/pg_dump/common.c b/src/bin/pg_dump/common.c index aa1589e3331d2..a1976fae607d6 100644 --- a/src/bin/pg_dump/common.c +++ b/src/bin/pg_dump/common.c @@ -17,6 +17,7 @@ #include +#include "catalog/pg_am_d.h" #include "catalog/pg_class_d.h" #include "catalog/pg_collation_d.h" #include "catalog/pg_extension_d.h" @@ -944,6 +945,24 @@ findOprByOid(Oid oid) return (OprInfo *) dobj; } +/* + * findAccessMethodByOid + * finds the DumpableObject for the access method with the given oid + * returns NULL if not found + */ +AccessMethodInfo * +findAccessMethodByOid(Oid oid) +{ + CatalogId catId; + DumpableObject *dobj; + + catId.tableoid = AccessMethodRelationId; + catId.oid = oid; + dobj = findObjectByCatalogId(catId); + Assert(dobj == NULL || dobj->objType == DO_ACCESS_METHOD); + return (AccessMethodInfo *) dobj; +} + /* * findCollationByOid * finds the DumpableObject for the collation with the given oid diff --git a/src/bin/pg_dump/meson.build b/src/bin/pg_dump/meson.build index 4a4ebbd8ec94f..a2233b0a1b431 100644 --- a/src/bin/pg_dump/meson.build +++ b/src/bin/pg_dump/meson.build @@ -102,7 +102,6 @@ tests += { 't/003_pg_dump_with_server.pl', 't/004_pg_dump_parallel.pl', 't/005_pg_dump_filterfile.pl', - 't/006_pg_dumpall.pl', 't/010_dump_connstr.pl', ], }, diff --git a/src/bin/pg_dump/parallel.c b/src/bin/pg_dump/parallel.c index 5974d6706fd57..086adcdc50295 100644 --- a/src/bin/pg_dump/parallel.c +++ b/src/bin/pg_dump/parallel.c @@ -333,16 +333,6 @@ on_exit_close_archive(Archive *AHX) on_exit_nicely(archive_close_connection, &shutdown_info); } -/* - * When pg_restore restores multiple databases, then update already added entry - * into array for cleanup. - */ -void -replace_on_exit_close_archive(Archive *AHX) -{ - shutdown_info.AHX = AHX; -} - /* * on_exit_nicely handler for shutting down database connections and * worker processes cleanly. diff --git a/src/bin/pg_dump/pg_backup.h b/src/bin/pg_dump/pg_backup.h index af0007fb6d2f1..4ebef1e864451 100644 --- a/src/bin/pg_dump/pg_backup.h +++ b/src/bin/pg_dump/pg_backup.h @@ -308,7 +308,7 @@ extern void SetArchiveOptions(Archive *AH, DumpOptions *dopt, RestoreOptions *ro extern void ProcessArchiveRestoreOptions(Archive *AHX); -extern void RestoreArchive(Archive *AHX, bool append_data); +extern void RestoreArchive(Archive *AHX); /* Open an existing archive */ extern Archive *OpenArchive(const char *FileSpec, const ArchiveFormat fmt); diff --git a/src/bin/pg_dump/pg_backup_archiver.c b/src/bin/pg_dump/pg_backup_archiver.c index 30e0da31aa340..dce88f040ace3 100644 --- a/src/bin/pg_dump/pg_backup_archiver.c +++ b/src/bin/pg_dump/pg_backup_archiver.c @@ -87,7 +87,7 @@ static int RestoringToDB(ArchiveHandle *AH); static void dump_lo_buf(ArchiveHandle *AH); static void dumpTimestamp(ArchiveHandle *AH, const char *msg, time_t tim); static void SetOutput(ArchiveHandle *AH, const char *filename, - const pg_compress_specification compression_spec, bool append_data); + const pg_compress_specification compression_spec); static CompressFileHandle *SaveOutput(ArchiveHandle *AH); static void RestoreOutput(ArchiveHandle *AH, CompressFileHandle *savedOutput); @@ -339,14 +339,9 @@ ProcessArchiveRestoreOptions(Archive *AHX) StrictNamesCheck(ropt); } -/* - * RestoreArchive - * - * If append_data is set, then append data into file as we are restoring dump - * of multiple databases which was taken by pg_dumpall. - */ +/* Public */ void -RestoreArchive(Archive *AHX, bool append_data) +RestoreArchive(Archive *AHX) { ArchiveHandle *AH = (ArchiveHandle *) AHX; RestoreOptions *ropt = AH->public.ropt; @@ -463,7 +458,7 @@ RestoreArchive(Archive *AHX, bool append_data) */ sav = SaveOutput(AH); if (ropt->filename || ropt->compression_spec.algorithm != PG_COMPRESSION_NONE) - SetOutput(AH, ropt->filename, ropt->compression_spec, append_data); + SetOutput(AH, ropt->filename, ropt->compression_spec); ahprintf(AH, "--\n-- PostgreSQL database dump\n--\n\n"); @@ -1302,7 +1297,7 @@ PrintTOCSummary(Archive *AHX) sav = SaveOutput(AH); if (ropt->filename) - SetOutput(AH, ropt->filename, out_compression_spec, false); + SetOutput(AH, ropt->filename, out_compression_spec); if (strftime(stamp_str, sizeof(stamp_str), PGDUMP_STRFTIME_FMT, localtime(&AH->createDate)) == 0) @@ -1681,8 +1676,7 @@ archprintf(Archive *AH, const char *fmt,...) static void SetOutput(ArchiveHandle *AH, const char *filename, - const pg_compress_specification compression_spec, - bool append_data) + const pg_compress_specification compression_spec) { CompressFileHandle *CFH; const char *mode; @@ -1702,7 +1696,7 @@ SetOutput(ArchiveHandle *AH, const char *filename, else fn = fileno(stdout); - if (append_data || AH->mode == archModeAppend) + if (AH->mode == archModeAppend) mode = PG_BINARY_A; else mode = PG_BINARY_W; diff --git a/src/bin/pg_dump/pg_backup_archiver.h b/src/bin/pg_dump/pg_backup_archiver.h index 365073b3eae45..325b53fc9bd4b 100644 --- a/src/bin/pg_dump/pg_backup_archiver.h +++ b/src/bin/pg_dump/pg_backup_archiver.h @@ -394,7 +394,6 @@ struct _tocEntry extern int parallel_restore(ArchiveHandle *AH, TocEntry *te); extern void on_exit_close_archive(Archive *AHX); -extern void replace_on_exit_close_archive(Archive *AHX); extern void warn_or_exit_horribly(ArchiveHandle *AH, const char *fmt,...) pg_attribute_printf(2, 3); diff --git a/src/bin/pg_dump/pg_backup_tar.c b/src/bin/pg_dump/pg_backup_tar.c index d94d0de2a5d17..b5ba3b46dd999 100644 --- a/src/bin/pg_dump/pg_backup_tar.c +++ b/src/bin/pg_dump/pg_backup_tar.c @@ -826,7 +826,7 @@ _CloseArchive(ArchiveHandle *AH) savVerbose = AH->public.verbose; AH->public.verbose = 0; - RestoreArchive((Archive *) AH, false); + RestoreArchive((Archive *) AH); SetArchiveOptions((Archive *) AH, savDopt, savRopt); diff --git a/src/bin/pg_dump/pg_dump.c b/src/bin/pg_dump/pg_dump.c index 6298edb26b5df..f3a353a61a58e 100644 --- a/src/bin/pg_dump/pg_dump.c +++ b/src/bin/pg_dump/pg_dump.c @@ -449,8 +449,6 @@ main(int argc, char **argv) bool data_only = false; bool schema_only = false; bool statistics_only = false; - bool with_data = false; - bool with_schema = false; bool with_statistics = false; bool no_data = false; bool no_schema = false; @@ -514,6 +512,7 @@ main(int argc, char **argv) {"section", required_argument, NULL, 5}, {"serializable-deferrable", no_argument, &dopt.serializable_deferrable, 1}, {"snapshot", required_argument, NULL, 6}, + {"statistics", no_argument, NULL, 22}, {"statistics-only", no_argument, NULL, 18}, {"strict-names", no_argument, &strict_names, 1}, {"use-set-session-authorization", no_argument, &dopt.use_setsessauth, 1}, @@ -528,9 +527,6 @@ main(int argc, char **argv) {"no-toast-compression", no_argument, &dopt.no_toast_compression, 1}, {"no-unlogged-table-data", no_argument, &dopt.no_unlogged_table_data, 1}, {"no-sync", no_argument, NULL, 7}, - {"with-data", no_argument, NULL, 22}, - {"with-schema", no_argument, NULL, 23}, - {"with-statistics", no_argument, NULL, 24}, {"on-conflict-do-nothing", no_argument, &dopt.do_nothing, 1}, {"rows-per-insert", required_argument, NULL, 10}, {"include-foreign-data", required_argument, NULL, 11}, @@ -798,14 +794,6 @@ main(int argc, char **argv) break; case 22: - with_data = true; - break; - - case 23: - with_schema = true; - break; - - case 24: with_statistics = true; break; @@ -852,13 +840,17 @@ main(int argc, char **argv) if (statistics_only && no_statistics) pg_fatal("options --statistics-only and --no-statistics cannot be used together"); - /* reject conflicting "with-" and "no-" options */ - if (with_data && no_data) - pg_fatal("options --with-data and --no-data cannot be used together"); - if (with_schema && no_schema) - pg_fatal("options --with-schema and --no-schema cannot be used together"); + /* reject conflicting "no-" options */ if (with_statistics && no_statistics) - pg_fatal("options --with-statistics and --no-statistics cannot be used together"); + pg_fatal("options --statistics and --no-statistics cannot be used together"); + + /* reject conflicting "-only" options */ + if (data_only && with_statistics) + pg_fatal("options %s and %s cannot be used together", + "-a/--data-only", "--statistics"); + if (schema_only && with_statistics) + pg_fatal("options %s and %s cannot be used together", + "-s/--schema-only", "--statistics"); if (schema_only && foreign_servers_include_patterns.head != NULL) pg_fatal("options -s/--schema-only and --include-foreign-data cannot be used together"); @@ -873,16 +865,14 @@ main(int argc, char **argv) pg_fatal("option --if-exists requires option -c/--clean"); /* - * Set derivative flags. An "-only" option may be overridden by an - * explicit "with-" option; e.g. "--schema-only --with-statistics" will - * include schema and statistics. Other ambiguous or nonsensical - * combinations, e.g. "--schema-only --no-schema", will have already - * caused an error in one of the checks above. + * Set derivative flags. Ambiguous or nonsensical combinations, e.g. + * "--schema-only --no-schema", will have already caused an error in one + * of the checks above. */ dopt.dumpData = ((dopt.dumpData && !schema_only && !statistics_only) || - (data_only || with_data)) && !no_data; + data_only) && !no_data; dopt.dumpSchema = ((dopt.dumpSchema && !data_only && !statistics_only) || - (schema_only || with_schema)) && !no_schema; + schema_only) && !no_schema; dopt.dumpStatistics = ((dopt.dumpStatistics && !schema_only && !data_only) || (statistics_only || with_statistics)) && !no_statistics; @@ -1265,7 +1255,7 @@ main(int argc, char **argv) * right now. */ if (plainText) - RestoreArchive(fout, false); + RestoreArchive(fout); CloseArchive(fout); @@ -1355,6 +1345,7 @@ help(const char *progname) printf(_(" --sequence-data include sequence data in dump\n")); printf(_(" --serializable-deferrable wait until the dump can run without anomalies\n")); printf(_(" --snapshot=SNAPSHOT use given snapshot for the dump\n")); + printf(_(" --statistics dump the statistics\n")); printf(_(" --statistics-only dump only the statistics, not schema or data\n")); printf(_(" --strict-names require table and/or schema include patterns to\n" " match at least one entity each\n")); @@ -1363,9 +1354,6 @@ help(const char *progname) printf(_(" --use-set-session-authorization\n" " use SET SESSION AUTHORIZATION commands instead of\n" " ALTER OWNER commands to set ownership\n")); - printf(_(" --with-data dump the data\n")); - printf(_(" --with-schema dump the schema\n")); - printf(_(" --with-statistics dump the statistics\n")); printf(_("\nConnection options:\n")); printf(_(" -d, --dbname=DBNAME database to dump\n")); @@ -2207,6 +2195,13 @@ selectDumpableProcLang(ProcLangInfo *plang, Archive *fout) static void selectDumpableAccessMethod(AccessMethodInfo *method, Archive *fout) { + /* see getAccessMethods() comment about v9.6. */ + if (fout->remoteVersion < 90600) + { + method->dobj.dump = DUMP_COMPONENT_NONE; + return; + } + if (checkExtensionMembership(&method->dobj, fout)) return; /* extension membership overrides all else */ @@ -6262,6 +6257,8 @@ getOperators(Archive *fout) int i_oprnamespace; int i_oprowner; int i_oprkind; + int i_oprleft; + int i_oprright; int i_oprcode; /* @@ -6273,6 +6270,8 @@ getOperators(Archive *fout) "oprnamespace, " "oprowner, " "oprkind, " + "oprleft, " + "oprright, " "oprcode::oid AS oprcode " "FROM pg_operator"); @@ -6288,6 +6287,8 @@ getOperators(Archive *fout) i_oprnamespace = PQfnumber(res, "oprnamespace"); i_oprowner = PQfnumber(res, "oprowner"); i_oprkind = PQfnumber(res, "oprkind"); + i_oprleft = PQfnumber(res, "oprleft"); + i_oprright = PQfnumber(res, "oprright"); i_oprcode = PQfnumber(res, "oprcode"); for (i = 0; i < ntups; i++) @@ -6301,6 +6302,8 @@ getOperators(Archive *fout) findNamespace(atooid(PQgetvalue(res, i, i_oprnamespace))); oprinfo[i].rolname = getRoleName(PQgetvalue(res, i, i_oprowner)); oprinfo[i].oprkind = (PQgetvalue(res, i, i_oprkind))[0]; + oprinfo[i].oprleft = atooid(PQgetvalue(res, i, i_oprleft)); + oprinfo[i].oprright = atooid(PQgetvalue(res, i, i_oprright)); oprinfo[i].oprcode = atooid(PQgetvalue(res, i, i_oprcode)); /* Decide whether we want to dump it */ @@ -6329,6 +6332,7 @@ getCollations(Archive *fout) int i_collname; int i_collnamespace; int i_collowner; + int i_collencoding; query = createPQExpBuffer(); @@ -6339,7 +6343,8 @@ getCollations(Archive *fout) appendPQExpBufferStr(query, "SELECT tableoid, oid, collname, " "collnamespace, " - "collowner " + "collowner, " + "collencoding " "FROM pg_collation"); res = ExecuteSqlQuery(fout, query->data, PGRES_TUPLES_OK); @@ -6353,6 +6358,7 @@ getCollations(Archive *fout) i_collname = PQfnumber(res, "collname"); i_collnamespace = PQfnumber(res, "collnamespace"); i_collowner = PQfnumber(res, "collowner"); + i_collencoding = PQfnumber(res, "collencoding"); for (i = 0; i < ntups; i++) { @@ -6364,6 +6370,7 @@ getCollations(Archive *fout) collinfo[i].dobj.namespace = findNamespace(atooid(PQgetvalue(res, i, i_collnamespace))); collinfo[i].rolname = getRoleName(PQgetvalue(res, i, i_collowner)); + collinfo[i].collencoding = atoi(PQgetvalue(res, i, i_collencoding)); /* Decide whether we want to dump it */ selectDumpableObject(&(collinfo[i].dobj), fout); @@ -6454,16 +6461,28 @@ getAccessMethods(Archive *fout) int i_amhandler; int i_amtype; - /* Before 9.6, there are no user-defined access methods */ - if (fout->remoteVersion < 90600) - return; - query = createPQExpBuffer(); - /* Select all access methods from pg_am table */ - appendPQExpBufferStr(query, "SELECT tableoid, oid, amname, amtype, " - "amhandler::pg_catalog.regproc AS amhandler " - "FROM pg_am"); + /* + * Select all access methods from pg_am table. v9.6 introduced CREATE + * ACCESS METHOD, so earlier versions usually have only built-in access + * methods. v9.6 also changed the access method API, replacing dozens of + * pg_am columns with amhandler. Even if a user created an access method + * by "INSERT INTO pg_am", we have no way to translate pre-v9.6 pg_am + * columns to a v9.6+ CREATE ACCESS METHOD. Hence, before v9.6, read + * pg_am just to facilitate findAccessMethodByOid() providing the + * OID-to-name mapping. + */ + appendPQExpBufferStr(query, "SELECT tableoid, oid, amname, "); + if (fout->remoteVersion >= 90600) + appendPQExpBufferStr(query, + "amtype, " + "amhandler::pg_catalog.regproc AS amhandler "); + else + appendPQExpBufferStr(query, + "'i'::pg_catalog.\"char\" AS amtype, " + "'-'::pg_catalog.regproc AS amhandler "); + appendPQExpBufferStr(query, "FROM pg_am"); res = ExecuteSqlQuery(fout, query->data, PGRES_TUPLES_OK); @@ -6512,6 +6531,7 @@ getOpclasses(Archive *fout) OpclassInfo *opcinfo; int i_tableoid; int i_oid; + int i_opcmethod; int i_opcname; int i_opcnamespace; int i_opcowner; @@ -6521,7 +6541,7 @@ getOpclasses(Archive *fout) * system-defined opclasses at dump-out time. */ - appendPQExpBufferStr(query, "SELECT tableoid, oid, opcname, " + appendPQExpBufferStr(query, "SELECT tableoid, oid, opcmethod, opcname, " "opcnamespace, " "opcowner " "FROM pg_opclass"); @@ -6534,6 +6554,7 @@ getOpclasses(Archive *fout) i_tableoid = PQfnumber(res, "tableoid"); i_oid = PQfnumber(res, "oid"); + i_opcmethod = PQfnumber(res, "opcmethod"); i_opcname = PQfnumber(res, "opcname"); i_opcnamespace = PQfnumber(res, "opcnamespace"); i_opcowner = PQfnumber(res, "opcowner"); @@ -6547,6 +6568,7 @@ getOpclasses(Archive *fout) opcinfo[i].dobj.name = pg_strdup(PQgetvalue(res, i, i_opcname)); opcinfo[i].dobj.namespace = findNamespace(atooid(PQgetvalue(res, i, i_opcnamespace))); + opcinfo[i].opcmethod = atooid(PQgetvalue(res, i, i_opcmethod)); opcinfo[i].rolname = getRoleName(PQgetvalue(res, i, i_opcowner)); /* Decide whether we want to dump it */ @@ -6572,6 +6594,7 @@ getOpfamilies(Archive *fout) OpfamilyInfo *opfinfo; int i_tableoid; int i_oid; + int i_opfmethod; int i_opfname; int i_opfnamespace; int i_opfowner; @@ -6583,7 +6606,7 @@ getOpfamilies(Archive *fout) * system-defined opfamilies at dump-out time. */ - appendPQExpBufferStr(query, "SELECT tableoid, oid, opfname, " + appendPQExpBufferStr(query, "SELECT tableoid, oid, opfmethod, opfname, " "opfnamespace, " "opfowner " "FROM pg_opfamily"); @@ -6597,6 +6620,7 @@ getOpfamilies(Archive *fout) i_tableoid = PQfnumber(res, "tableoid"); i_oid = PQfnumber(res, "oid"); i_opfname = PQfnumber(res, "opfname"); + i_opfmethod = PQfnumber(res, "opfmethod"); i_opfnamespace = PQfnumber(res, "opfnamespace"); i_opfowner = PQfnumber(res, "opfowner"); @@ -6609,6 +6633,7 @@ getOpfamilies(Archive *fout) opfinfo[i].dobj.name = pg_strdup(PQgetvalue(res, i, i_opfname)); opfinfo[i].dobj.namespace = findNamespace(atooid(PQgetvalue(res, i, i_opfnamespace))); + opfinfo[i].opfmethod = atooid(PQgetvalue(res, i, i_opfmethod)); opfinfo[i].rolname = getRoleName(PQgetvalue(res, i, i_opfowner)); /* Decide whether we want to dump it */ diff --git a/src/bin/pg_dump/pg_dump.h b/src/bin/pg_dump/pg_dump.h index 93a4475d51b80..dde85ed156cc8 100644 --- a/src/bin/pg_dump/pg_dump.h +++ b/src/bin/pg_dump/pg_dump.h @@ -260,6 +260,8 @@ typedef struct _oprInfo DumpableObject dobj; const char *rolname; char oprkind; + Oid oprleft; + Oid oprright; Oid oprcode; } OprInfo; @@ -273,12 +275,14 @@ typedef struct _accessMethodInfo typedef struct _opclassInfo { DumpableObject dobj; + Oid opcmethod; const char *rolname; } OpclassInfo; typedef struct _opfamilyInfo { DumpableObject dobj; + Oid opfmethod; const char *rolname; } OpfamilyInfo; @@ -286,6 +290,7 @@ typedef struct _collInfo { DumpableObject dobj; const char *rolname; + int collencoding; } CollInfo; typedef struct _convInfo @@ -760,6 +765,7 @@ extern TableInfo *findTableByOid(Oid oid); extern TypeInfo *findTypeByOid(Oid oid); extern FuncInfo *findFuncByOid(Oid oid); extern OprInfo *findOprByOid(Oid oid); +extern AccessMethodInfo *findAccessMethodByOid(Oid oid); extern CollInfo *findCollationByOid(Oid oid); extern NamespaceInfo *findNamespaceByOid(Oid oid); extern ExtensionInfo *findExtensionByOid(Oid oid); diff --git a/src/bin/pg_dump/pg_dump_sort.c b/src/bin/pg_dump/pg_dump_sort.c index f99a0797ea7fb..a02da3e9652c1 100644 --- a/src/bin/pg_dump/pg_dump_sort.c +++ b/src/bin/pg_dump/pg_dump_sort.c @@ -162,6 +162,8 @@ static DumpId postDataBoundId; static int DOTypeNameCompare(const void *p1, const void *p2); +static int pgTypeNameCompare(Oid typid1, Oid typid2); +static int accessMethodNameCompare(Oid am1, Oid am2); static bool TopoSort(DumpableObject **objs, int numObjs, DumpableObject **ordering, @@ -228,12 +230,39 @@ DOTypeNameCompare(const void *p1, const void *p2) else if (obj2->namespace) return 1; - /* Sort by name */ + /* + * Sort by name. With a few exceptions, names here are single catalog + * columns. To get a fuller picture, grep pg_dump.c for "dobj.name = ". + * Names here don't match "Name:" in plain format output, which is a + * _tocEntry.tag. For example, DumpableObject.name of a constraint is + * pg_constraint.conname, but _tocEntry.tag of a constraint is relname and + * conname joined with a space. + */ cmpval = strcmp(obj1->name, obj2->name); if (cmpval != 0) return cmpval; - /* To have a stable sort order, break ties for some object types */ + /* + * Sort by type. This helps types that share a type priority without + * sharing a unique name constraint, e.g. opclass and opfamily. + */ + cmpval = obj1->objType - obj2->objType; + if (cmpval != 0) + return cmpval; + + /* + * To have a stable sort order, break ties for some object types. Most + * catalogs have a natural key, e.g. pg_proc_proname_args_nsp_index. Where + * the above "namespace" and "name" comparisons don't cover all natural + * key columns, compare the rest here. + * + * The natural key usually refers to other catalogs by surrogate keys. + * Hence, this translates each of those references to the natural key of + * the referenced catalog. That may descend through multiple levels of + * catalog references. For example, to sort by pg_proc.proargtypes, + * descend to each pg_type and then further to its pg_namespace, for an + * overall sort by (nspname, typname). + */ if (obj1->objType == DO_FUNC || obj1->objType == DO_AGG) { FuncInfo *fobj1 = *(FuncInfo *const *) p1; @@ -246,22 +275,10 @@ DOTypeNameCompare(const void *p1, const void *p2) return cmpval; for (i = 0; i < fobj1->nargs; i++) { - TypeInfo *argtype1 = findTypeByOid(fobj1->argtypes[i]); - TypeInfo *argtype2 = findTypeByOid(fobj2->argtypes[i]); - - if (argtype1 && argtype2) - { - if (argtype1->dobj.namespace && argtype2->dobj.namespace) - { - cmpval = strcmp(argtype1->dobj.namespace->dobj.name, - argtype2->dobj.namespace->dobj.name); - if (cmpval != 0) - return cmpval; - } - cmpval = strcmp(argtype1->dobj.name, argtype2->dobj.name); - if (cmpval != 0) - return cmpval; - } + cmpval = pgTypeNameCompare(fobj1->argtypes[i], + fobj2->argtypes[i]); + if (cmpval != 0) + return cmpval; } } else if (obj1->objType == DO_OPERATOR) @@ -273,6 +290,57 @@ DOTypeNameCompare(const void *p1, const void *p2) cmpval = (oobj2->oprkind - oobj1->oprkind); if (cmpval != 0) return cmpval; + /* Within an oprkind, sort by argument type names */ + cmpval = pgTypeNameCompare(oobj1->oprleft, oobj2->oprleft); + if (cmpval != 0) + return cmpval; + cmpval = pgTypeNameCompare(oobj1->oprright, oobj2->oprright); + if (cmpval != 0) + return cmpval; + } + else if (obj1->objType == DO_OPCLASS) + { + OpclassInfo *opcobj1 = *(OpclassInfo *const *) p1; + OpclassInfo *opcobj2 = *(OpclassInfo *const *) p2; + + /* Sort by access method name, per pg_opclass_am_name_nsp_index */ + cmpval = accessMethodNameCompare(opcobj1->opcmethod, + opcobj2->opcmethod); + if (cmpval != 0) + return cmpval; + } + else if (obj1->objType == DO_OPFAMILY) + { + OpfamilyInfo *opfobj1 = *(OpfamilyInfo *const *) p1; + OpfamilyInfo *opfobj2 = *(OpfamilyInfo *const *) p2; + + /* Sort by access method name, per pg_opfamily_am_name_nsp_index */ + cmpval = accessMethodNameCompare(opfobj1->opfmethod, + opfobj2->opfmethod); + if (cmpval != 0) + return cmpval; + } + else if (obj1->objType == DO_COLLATION) + { + CollInfo *cobj1 = *(CollInfo *const *) p1; + CollInfo *cobj2 = *(CollInfo *const *) p2; + + /* + * Sort by encoding, per pg_collation_name_enc_nsp_index. Technically, + * this is not necessary, because wherever this changes dump order, + * restoring the dump fails anyway. CREATE COLLATION can't create a + * tie for this to break, because it imposes restrictions to make + * (nspname, collname) uniquely identify a collation within a given + * DatabaseEncoding. While pg_import_system_collations() can create a + * tie, pg_dump+restore fails after + * pg_import_system_collations('my_schema') does so. However, there's + * little to gain by ignoring one natural key column on the basis of + * those limitations elsewhere, so respect the full natural key like + * we do for other object types. + */ + cmpval = cobj1->collencoding - cobj2->collencoding; + if (cmpval != 0) + return cmpval; } else if (obj1->objType == DO_ATTRDEF) { @@ -317,11 +385,143 @@ DOTypeNameCompare(const void *p1, const void *p2) if (cmpval != 0) return cmpval; } + else if (obj1->objType == DO_CONSTRAINT) + { + ConstraintInfo *robj1 = *(ConstraintInfo *const *) p1; + ConstraintInfo *robj2 = *(ConstraintInfo *const *) p2; - /* Usually shouldn't get here, but if we do, sort by OID */ + /* + * Sort domain constraints before table constraints, for consistency + * with our decision to sort CREATE DOMAIN before CREATE TABLE. + */ + if (robj1->condomain) + { + if (robj2->condomain) + { + /* Sort by domain name (domain namespace was considered) */ + cmpval = strcmp(robj1->condomain->dobj.name, + robj2->condomain->dobj.name); + if (cmpval != 0) + return cmpval; + } + else + return PRIO_TYPE - PRIO_TABLE; + } + else if (robj2->condomain) + return PRIO_TABLE - PRIO_TYPE; + else + { + /* Sort by table name (table namespace was considered already) */ + cmpval = strcmp(robj1->contable->dobj.name, + robj2->contable->dobj.name); + if (cmpval != 0) + return cmpval; + } + } + else if (obj1->objType == DO_PUBLICATION_REL) + { + PublicationRelInfo *probj1 = *(PublicationRelInfo *const *) p1; + PublicationRelInfo *probj2 = *(PublicationRelInfo *const *) p2; + + /* Sort by publication name, since (namespace, name) match the rel */ + cmpval = strcmp(probj1->publication->dobj.name, + probj2->publication->dobj.name); + if (cmpval != 0) + return cmpval; + } + else if (obj1->objType == DO_PUBLICATION_TABLE_IN_SCHEMA) + { + PublicationSchemaInfo *psobj1 = *(PublicationSchemaInfo *const *) p1; + PublicationSchemaInfo *psobj2 = *(PublicationSchemaInfo *const *) p2; + + /* Sort by publication name, since ->name is just nspname */ + cmpval = strcmp(psobj1->publication->dobj.name, + psobj2->publication->dobj.name); + if (cmpval != 0) + return cmpval; + } + + /* + * Shouldn't get here except after catalog corruption, but if we do, sort + * by OID. This may make logically-identical databases differ in the + * order of objects in dump output. Users will get spurious schema diffs. + * Expect flaky failures of 002_pg_upgrade.pl test 'dump outputs from + * original and restored regression databases match' if the regression + * database contains objects allowing that test to reach here. That's a + * consequence of the test using "pg_restore -j", which doesn't fully + * constrain OID assignment order. + */ + Assert(false); return oidcmp(obj1->catId.oid, obj2->catId.oid); } +/* Compare two OID-identified pg_type values by nspname, then by typname. */ +static int +pgTypeNameCompare(Oid typid1, Oid typid2) +{ + TypeInfo *typobj1; + TypeInfo *typobj2; + int cmpval; + + if (typid1 == typid2) + return 0; + + typobj1 = findTypeByOid(typid1); + typobj2 = findTypeByOid(typid2); + + if (!typobj1 || !typobj2) + { + /* + * getTypes() didn't find some OID. Assume catalog corruption, e.g. + * an oprright value without the corresponding OID in a pg_type row. + * Report as "equal", so the caller uses the next available basis for + * comparison, e.g. the next function argument. + * + * Unary operators have InvalidOid in oprleft (if oprkind='r') or in + * oprright (if oprkind='l'). Caller already sorted by oprkind, + * calling us only for like-kind operators. Hence, "typid1 == typid2" + * took care of InvalidOid. (v14 removed postfix operator support. + * Hence, when dumping from v14+, only oprleft can be InvalidOid.) + */ + Assert(false); + return 0; + } + + if (!typobj1->dobj.namespace || !typobj2->dobj.namespace) + Assert(false); /* catalog corruption */ + else + { + cmpval = strcmp(typobj1->dobj.namespace->dobj.name, + typobj2->dobj.namespace->dobj.name); + if (cmpval != 0) + return cmpval; + } + return strcmp(typobj1->dobj.name, typobj2->dobj.name); +} + +/* Compare two OID-identified pg_am values by amname. */ +static int +accessMethodNameCompare(Oid am1, Oid am2) +{ + AccessMethodInfo *amobj1; + AccessMethodInfo *amobj2; + + if (am1 == am2) + return 0; + + amobj1 = findAccessMethodByOid(am1); + amobj2 = findAccessMethodByOid(am2); + + if (!amobj1 || !amobj2) + { + /* catalog corruption: handle like pgTypeNameCompare() does */ + Assert(false); + return 0; + } + + return strcmp(amobj1->dobj.name, amobj2->dobj.name); +} + /* * Sort the given objects into a safe dump order using dependency diff --git a/src/bin/pg_dump/pg_dumpall.c b/src/bin/pg_dump/pg_dumpall.c index 100317b1aa949..27aa1b656989c 100644 --- a/src/bin/pg_dump/pg_dumpall.c +++ b/src/bin/pg_dump/pg_dumpall.c @@ -65,10 +65,9 @@ static void dropTablespaces(PGconn *conn); static void dumpTablespaces(PGconn *conn); static void dropDBs(PGconn *conn); static void dumpUserConfig(PGconn *conn, const char *username); -static void dumpDatabases(PGconn *conn, ArchiveFormat archDumpFormat); +static void dumpDatabases(PGconn *conn); static void dumpTimestamp(const char *msg); -static int runPgDump(const char *dbname, const char *create_opts, - char *dbfile, ArchiveFormat archDumpFormat); +static int runPgDump(const char *dbname, const char *create_opts); static void buildShSecLabels(PGconn *conn, const char *catalog_name, Oid objectId, const char *objtype, const char *objname, @@ -77,7 +76,6 @@ static void executeCommand(PGconn *conn, const char *query); static void expand_dbname_patterns(PGconn *conn, SimpleStringList *patterns, SimpleStringList *names); static void read_dumpall_filters(const char *filename, SimpleStringList *pattern); -static ArchiveFormat parseDumpFormat(const char *format); static char pg_dump_bin[MAXPGPATH]; static PQExpBuffer pgdumpopts; @@ -107,8 +105,6 @@ static int no_subscriptions = 0; static int no_toast_compression = 0; static int no_unlogged_table_data = 0; static int no_role_passwords = 0; -static int with_data = 0; -static int with_schema = 0; static int with_statistics = 0; static int server_version; static int load_via_partition_root = 0; @@ -150,7 +146,6 @@ main(int argc, char *argv[]) {"password", no_argument, NULL, 'W'}, {"no-privileges", no_argument, NULL, 'x'}, {"no-acl", no_argument, NULL, 'x'}, - {"format", required_argument, NULL, 'F'}, /* * the following options don't have an equivalent short option letter @@ -183,11 +178,9 @@ main(int argc, char *argv[]) {"no-sync", no_argument, NULL, 4}, {"no-toast-compression", no_argument, &no_toast_compression, 1}, {"no-unlogged-table-data", no_argument, &no_unlogged_table_data, 1}, - {"with-data", no_argument, &with_data, 1}, - {"with-schema", no_argument, &with_schema, 1}, - {"with-statistics", no_argument, &with_statistics, 1}, {"on-conflict-do-nothing", no_argument, &on_conflict_do_nothing, 1}, {"rows-per-insert", required_argument, NULL, 7}, + {"statistics", no_argument, &with_statistics, 1}, {"statistics-only", no_argument, &statistics_only, 1}, {"filter", required_argument, NULL, 8}, {"sequence-data", no_argument, &sequence_data, 1}, @@ -201,8 +194,6 @@ main(int argc, char *argv[]) char *pgdb = NULL; char *use_role = NULL; const char *dumpencoding = NULL; - ArchiveFormat archDumpFormat = archNull; - const char *formatName = "p"; trivalue prompt_password = TRI_DEFAULT; bool data_only = false; bool globals_only = false; @@ -252,7 +243,7 @@ main(int argc, char *argv[]) pgdumpopts = createPQExpBuffer(); - while ((c = getopt_long(argc, argv, "acd:E:f:F:gh:l:Op:rsS:tU:vwWx", long_options, &optindex)) != -1) + while ((c = getopt_long(argc, argv, "acd:E:f:gh:l:Op:rsS:tU:vwWx", long_options, &optindex)) != -1) { switch (c) { @@ -280,9 +271,7 @@ main(int argc, char *argv[]) appendPQExpBufferStr(pgdumpopts, " -f "); appendShellString(pgdumpopts, filename); break; - case 'F': - formatName = pg_strdup(optarg); - break; + case 'g': globals_only = true; break; @@ -431,21 +420,6 @@ main(int argc, char *argv[]) exit_nicely(1); } - /* Get format for dump. */ - archDumpFormat = parseDumpFormat(formatName); - - /* - * If a non-plain format is specified, a file name is also required as the - * path to the main directory. - */ - if (archDumpFormat != archNull && - (!filename || strcmp(filename, "") == 0)) - { - pg_log_error("option -F/--format=d|c|t requires option -f/--file"); - pg_log_error_hint("Try \"%s --help\" for more information.", progname); - exit_nicely(1); - } - /* * If password values are not required in the dump, switch to using * pg_roles which is equally useful, just more likely to have unrestricted @@ -497,12 +471,8 @@ main(int argc, char *argv[]) appendPQExpBufferStr(pgdumpopts, " --no-toast-compression"); if (no_unlogged_table_data) appendPQExpBufferStr(pgdumpopts, " --no-unlogged-table-data"); - if (with_data) - appendPQExpBufferStr(pgdumpopts, " --with-data"); - if (with_schema) - appendPQExpBufferStr(pgdumpopts, " --with-schema"); if (with_statistics) - appendPQExpBufferStr(pgdumpopts, " --with-statistics"); + appendPQExpBufferStr(pgdumpopts, " --statistics"); if (on_conflict_do_nothing) appendPQExpBufferStr(pgdumpopts, " --on-conflict-do-nothing"); if (statistics_only) @@ -510,33 +480,6 @@ main(int argc, char *argv[]) if (sequence_data) appendPQExpBufferStr(pgdumpopts, " --sequence-data"); - /* - * Open the output file if required, otherwise use stdout. If required, - * then create new directory and global.dat file. - */ - if (archDumpFormat != archNull) - { - char global_path[MAXPGPATH]; - - /* Create new directory or accept the empty existing directory. */ - create_or_open_dir(filename); - - snprintf(global_path, MAXPGPATH, "%s/global.dat", filename); - - OPF = fopen(global_path, PG_BINARY_W); - if (!OPF) - pg_fatal("could not open file \"%s\": %m", global_path); - } - else if (filename) - { - OPF = fopen(filename, PG_BINARY_W); - if (!OPF) - pg_fatal("could not open output file \"%s\": %m", - filename); - } - else - OPF = stdout; - /* * If there was a database specified on the command line, use that, * otherwise try to connect to database "postgres", and failing that @@ -576,6 +519,19 @@ main(int argc, char *argv[]) expand_dbname_patterns(conn, &database_exclude_patterns, &database_exclude_names); + /* + * Open the output file if required, otherwise use stdout + */ + if (filename) + { + OPF = fopen(filename, PG_BINARY_W); + if (!OPF) + pg_fatal("could not open output file \"%s\": %m", + filename); + } + else + OPF = stdout; + /* * Set the client encoding if requested. */ @@ -675,7 +631,7 @@ main(int argc, char *argv[]) } if (!globals_only && !roles_only && !tablespaces_only) - dumpDatabases(conn, archDumpFormat); + dumpDatabases(conn); PQfinish(conn); @@ -688,7 +644,7 @@ main(int argc, char *argv[]) fclose(OPF); /* sync the resulting file, errors are not fatal */ - if (dosync && (archDumpFormat == archNull)) + if (dosync) (void) fsync_fname(filename, false); } @@ -699,14 +655,12 @@ main(int argc, char *argv[]) static void help(void) { - printf(_("%s exports a PostgreSQL database cluster as an SQL script or to other formats.\n\n"), progname); + printf(_("%s exports a PostgreSQL database cluster as an SQL script.\n\n"), progname); printf(_("Usage:\n")); printf(_(" %s [OPTION]...\n"), progname); printf(_("\nGeneral options:\n")); printf(_(" -f, --file=FILENAME output file name\n")); - printf(_(" -F, --format=c|d|t|p output file format (custom, directory, tar,\n" - " plain text (default))\n")); printf(_(" -v, --verbose verbose mode\n")); printf(_(" -V, --version output version information, then exit\n")); printf(_(" --lock-wait-timeout=TIMEOUT fail after waiting TIMEOUT for a table lock\n")); @@ -750,13 +704,11 @@ help(void) printf(_(" --quote-all-identifiers quote all identifiers, even if not key words\n")); printf(_(" --rows-per-insert=NROWS number of rows per INSERT; implies --inserts\n")); printf(_(" --sequence-data include sequence data in dump\n")); + printf(_(" --statistics dump the statistics\n")); printf(_(" --statistics-only dump only the statistics, not schema or data\n")); printf(_(" --use-set-session-authorization\n" " use SET SESSION AUTHORIZATION commands instead of\n" " ALTER OWNER commands to set ownership\n")); - printf(_(" --with-data dump the data\n")); - printf(_(" --with-schema dump the schema\n")); - printf(_(" --with-statistics dump the statistics\n")); printf(_("\nConnection options:\n")); printf(_(" -d, --dbname=CONNSTR connect using connection string\n")); @@ -1013,6 +965,9 @@ dumpRoles(PGconn *conn) * We do it this way because config settings for roles could mention the * names of other roles. */ + if (PQntuples(res) > 0) + fprintf(OPF, "\n--\n-- User Configurations\n--\n"); + for (i = 0; i < PQntuples(res); i++) dumpUserConfig(conn, PQgetvalue(res, i, i_rolname)); @@ -1526,7 +1481,6 @@ dumpUserConfig(PGconn *conn, const char *username) { PQExpBuffer buf = createPQExpBuffer(); PGresult *res; - static bool header_done = false; printfPQExpBuffer(buf, "SELECT unnest(setconfig) FROM pg_db_role_setting " "WHERE setdatabase = 0 AND setrole = " @@ -1538,13 +1492,7 @@ dumpUserConfig(PGconn *conn, const char *username) res = executeQuery(conn, buf->data); if (PQntuples(res) > 0) - { - if (!header_done) - fprintf(OPF, "\n--\n-- User Configurations\n--\n"); - header_done = true; - fprintf(OPF, "\n--\n-- User Config \"%s\"\n--\n\n", username); - } for (int i = 0; i < PQntuples(res); i++) { @@ -1618,13 +1566,10 @@ expand_dbname_patterns(PGconn *conn, * Dump contents of databases. */ static void -dumpDatabases(PGconn *conn, ArchiveFormat archDumpFormat) +dumpDatabases(PGconn *conn) { PGresult *res; int i; - char db_subdir[MAXPGPATH]; - char dbfilepath[MAXPGPATH]; - FILE *map_file = NULL; /* * Skip databases marked not datallowconn, since we'd be unable to connect @@ -1638,42 +1583,18 @@ dumpDatabases(PGconn *conn, ArchiveFormat archDumpFormat) * doesn't have some failure mode with --clean. */ res = executeQuery(conn, - "SELECT datname, oid " + "SELECT datname " "FROM pg_database d " "WHERE datallowconn AND datconnlimit != -2 " "ORDER BY (datname <> 'template1'), datname"); - if (archDumpFormat == archNull && PQntuples(res) > 0) + if (PQntuples(res) > 0) fprintf(OPF, "--\n-- Databases\n--\n\n"); - /* - * If directory/tar/custom format is specified, create a subdirectory - * under the main directory and each database dump file or subdirectory - * will be created in that subdirectory by pg_dump. - */ - if (archDumpFormat != archNull) - { - char map_file_path[MAXPGPATH]; - - snprintf(db_subdir, MAXPGPATH, "%s/databases", filename); - - /* Create a subdirectory with 'databases' name under main directory. */ - if (mkdir(db_subdir, pg_dir_create_mode) != 0) - pg_fatal("could not create directory \"%s\": %m", db_subdir); - - snprintf(map_file_path, MAXPGPATH, "%s/map.dat", filename); - - /* Create a map file (to store dboid and dbname) */ - map_file = fopen(map_file_path, PG_BINARY_W); - if (!map_file) - pg_fatal("could not open file \"%s\": %m", map_file_path); - } - for (i = 0; i < PQntuples(res); i++) { char *dbname = PQgetvalue(res, i, 0); - char *oid = PQgetvalue(res, i, 1); - const char *create_opts = ""; + const char *create_opts; int ret; /* Skip template0, even if it's not marked !datallowconn. */ @@ -1687,27 +1608,9 @@ dumpDatabases(PGconn *conn, ArchiveFormat archDumpFormat) continue; } - /* - * If this is not a plain format dump, then append dboid and dbname to - * the map.dat file. - */ - if (archDumpFormat != archNull) - { - if (archDumpFormat == archCustom) - snprintf(dbfilepath, MAXPGPATH, "\"%s\"/\"%s\".dmp", db_subdir, oid); - else if (archDumpFormat == archTar) - snprintf(dbfilepath, MAXPGPATH, "\"%s\"/\"%s\".tar", db_subdir, oid); - else - snprintf(dbfilepath, MAXPGPATH, "\"%s\"/\"%s\"", db_subdir, oid); - - /* Put one line entry for dboid and dbname in map file. */ - fprintf(map_file, "%s %s\n", oid, dbname); - } - pg_log_info("dumping database \"%s\"", dbname); - if (archDumpFormat == archNull) - fprintf(OPF, "--\n-- Database \"%s\" dump\n--\n\n", dbname); + fprintf(OPF, "--\n-- Database \"%s\" dump\n--\n\n", dbname); /* * We assume that "template1" and "postgres" already exist in the @@ -1721,9 +1624,12 @@ dumpDatabases(PGconn *conn, ArchiveFormat archDumpFormat) { if (output_clean) create_opts = "--clean --create"; - /* Since pg_dump won't emit a \connect command, we must */ - else if (archDumpFormat == archNull) + else + { + create_opts = ""; + /* Since pg_dump won't emit a \connect command, we must */ fprintf(OPF, "\\connect %s\n\n", dbname); + } } else create_opts = "--create"; @@ -1731,30 +1637,19 @@ dumpDatabases(PGconn *conn, ArchiveFormat archDumpFormat) if (filename) fclose(OPF); - ret = runPgDump(dbname, create_opts, dbfilepath, archDumpFormat); + ret = runPgDump(dbname, create_opts); if (ret != 0) pg_fatal("pg_dump failed on database \"%s\", exiting", dbname); if (filename) { - char global_path[MAXPGPATH]; - - if (archDumpFormat != archNull) - snprintf(global_path, MAXPGPATH, "%s/global.dat", filename); - else - snprintf(global_path, MAXPGPATH, "%s", filename); - - OPF = fopen(global_path, PG_BINARY_A); + OPF = fopen(filename, PG_BINARY_A); if (!OPF) pg_fatal("could not re-open the output file \"%s\": %m", - global_path); + filename); } } - /* Close map file */ - if (archDumpFormat != archNull) - fclose(map_file); - PQclear(res); } @@ -1764,8 +1659,7 @@ dumpDatabases(PGconn *conn, ArchiveFormat archDumpFormat) * Run pg_dump on dbname, with specified options. */ static int -runPgDump(const char *dbname, const char *create_opts, char *dbfile, - ArchiveFormat archDumpFormat) +runPgDump(const char *dbname, const char *create_opts) { PQExpBufferData connstrbuf; PQExpBufferData cmd; @@ -1774,36 +1668,17 @@ runPgDump(const char *dbname, const char *create_opts, char *dbfile, initPQExpBuffer(&connstrbuf); initPQExpBuffer(&cmd); + printfPQExpBuffer(&cmd, "\"%s\" %s %s", pg_dump_bin, + pgdumpopts->data, create_opts); + /* - * If this is not a plain format dump, then append file name and dump - * format to the pg_dump command to get archive dump. + * If we have a filename, use the undocumented plain-append pg_dump + * format. */ - if (archDumpFormat != archNull) - { - printfPQExpBuffer(&cmd, "\"%s\" -f %s %s", pg_dump_bin, - dbfile, create_opts); - - if (archDumpFormat == archDirectory) - appendPQExpBufferStr(&cmd, " --format=directory "); - else if (archDumpFormat == archCustom) - appendPQExpBufferStr(&cmd, " --format=custom "); - else if (archDumpFormat == archTar) - appendPQExpBufferStr(&cmd, " --format=tar "); - } + if (filename) + appendPQExpBufferStr(&cmd, " -Fa "); else - { - printfPQExpBuffer(&cmd, "\"%s\" %s %s", pg_dump_bin, - pgdumpopts->data, create_opts); - - /* - * If we have a filename, use the undocumented plain-append pg_dump - * format. - */ - if (filename) - appendPQExpBufferStr(&cmd, " -Fa "); - else - appendPQExpBufferStr(&cmd, " -Fp "); - } + appendPQExpBufferStr(&cmd, " -Fp "); /* * Append the database name to the already-constructed stem of connection @@ -1948,36 +1823,3 @@ read_dumpall_filters(const char *filename, SimpleStringList *pattern) filter_free(&fstate); } - -/* - * parseDumpFormat - * - * This will validate dump formats. - */ -static ArchiveFormat -parseDumpFormat(const char *format) -{ - ArchiveFormat archDumpFormat; - - if (pg_strcasecmp(format, "c") == 0) - archDumpFormat = archCustom; - else if (pg_strcasecmp(format, "custom") == 0) - archDumpFormat = archCustom; - else if (pg_strcasecmp(format, "d") == 0) - archDumpFormat = archDirectory; - else if (pg_strcasecmp(format, "directory") == 0) - archDumpFormat = archDirectory; - else if (pg_strcasecmp(format, "p") == 0) - archDumpFormat = archNull; - else if (pg_strcasecmp(format, "plain") == 0) - archDumpFormat = archNull; - else if (pg_strcasecmp(format, "t") == 0) - archDumpFormat = archTar; - else if (pg_strcasecmp(format, "tar") == 0) - archDumpFormat = archTar; - else - pg_fatal("unrecognized output format \"%s\"; please specify \"c\", \"d\", \"p\", or \"t\"", - format); - - return archDumpFormat; -} diff --git a/src/bin/pg_dump/pg_restore.c b/src/bin/pg_dump/pg_restore.c index 6ef789cb06d63..6c129278bc52b 100644 --- a/src/bin/pg_dump/pg_restore.c +++ b/src/bin/pg_dump/pg_restore.c @@ -2,7 +2,7 @@ * * pg_restore.c * pg_restore is an utility extracting postgres database definitions - * from a backup archive created by pg_dump/pg_dumpall using the archiver + * from a backup archive created by pg_dump using the archiver * interface. * * pg_restore will read the backup archive and @@ -41,15 +41,11 @@ #include "postgres_fe.h" #include -#include #ifdef HAVE_TERMIOS_H #include #endif -#include "common/string.h" -#include "connectdb.h" #include "fe_utils/option_utils.h" -#include "fe_utils/string_utils.h" #include "filter.h" #include "getopt_long.h" #include "parallel.h" @@ -57,43 +53,18 @@ static void usage(const char *progname); static void read_restore_filters(const char *filename, RestoreOptions *opts); -static bool file_exists_in_directory(const char *dir, const char *filename); -static int restore_one_database(const char *inputFileSpec, RestoreOptions *opts, - int numWorkers, bool append_data, int num); -static int read_one_statement(StringInfo inBuf, FILE *pfile); -static int restore_all_databases(PGconn *conn, const char *dumpdirpath, - SimpleStringList db_exclude_patterns, RestoreOptions *opts, int numWorkers); -static int process_global_sql_commands(PGconn *conn, const char *dumpdirpath, - const char *outfile); -static void copy_or_print_global_file(const char *outfile, FILE *pfile); -static int get_dbnames_list_to_restore(PGconn *conn, - SimplePtrList *dbname_oid_list, - SimpleStringList db_exclude_patterns); -static int get_dbname_oid_list_from_mfile(const char *dumpdirpath, - SimplePtrList *dbname_oid_list); - -/* - * Stores a database OID and the corresponding name. - */ -typedef struct DbOidName -{ - Oid oid; - char str[FLEXIBLE_ARRAY_MEMBER]; /* null-terminated string here */ -} DbOidName; - int main(int argc, char **argv) { RestoreOptions *opts; int c; + int exit_code; int numWorkers = 1; + Archive *AH; char *inputFileSpec; bool data_only = false; bool schema_only = false; - int n_errors = 0; - bool globals_only = false; - SimpleStringList db_exclude_patterns = {NULL, NULL}; static int disable_triggers = 0; static int enable_row_security = 0; static int if_exists = 0; @@ -111,15 +82,12 @@ main(int argc, char **argv) static int no_subscriptions = 0; static int strict_names = 0; static int statistics_only = 0; - static int with_data = 0; - static int with_schema = 0; static int with_statistics = 0; struct option cmdopts[] = { {"clean", 0, NULL, 'c'}, {"create", 0, NULL, 'C'}, {"data-only", 0, NULL, 'a'}, - {"globals-only", 0, NULL, 'g'}, {"dbname", 1, NULL, 'd'}, {"exit-on-error", 0, NULL, 'e'}, {"exclude-schema", 1, NULL, 'N'}, @@ -169,12 +137,9 @@ main(int argc, char **argv) {"no-security-labels", no_argument, &no_security_labels, 1}, {"no-subscriptions", no_argument, &no_subscriptions, 1}, {"no-statistics", no_argument, &no_statistics, 1}, - {"with-data", no_argument, &with_data, 1}, - {"with-schema", no_argument, &with_schema, 1}, - {"with-statistics", no_argument, &with_statistics, 1}, + {"statistics", no_argument, &with_statistics, 1}, {"statistics-only", no_argument, &statistics_only, 1}, {"filter", required_argument, NULL, 4}, - {"exclude-database", required_argument, NULL, 6}, {NULL, 0, NULL, 0} }; @@ -203,7 +168,7 @@ main(int argc, char **argv) } } - while ((c = getopt_long(argc, argv, "acCd:ef:F:gh:I:j:lL:n:N:Op:P:RsS:t:T:U:vwWx1", + while ((c = getopt_long(argc, argv, "acCd:ef:F:h:I:j:lL:n:N:Op:P:RsS:t:T:U:vwWx1", cmdopts, NULL)) != -1) { switch (c) @@ -230,14 +195,11 @@ main(int argc, char **argv) if (strlen(optarg) != 0) opts->formatName = pg_strdup(optarg); break; - case 'g': - /* restore only global.dat file from directory */ - globals_only = true; - break; case 'h': if (strlen(optarg) != 0) opts->cparams.pghost = pg_strdup(optarg); break; + case 'j': /* number of restore jobs */ if (!option_parse_int(optarg, "-j/--jobs", 1, PG_MAX_JOBS, @@ -352,9 +314,6 @@ main(int argc, char **argv) exit(1); opts->exit_on_error = true; break; - case 6: /* database patterns to skip */ - simple_string_list_append(&db_exclude_patterns, optarg); - break; default: /* getopt_long already emitted a complaint */ @@ -382,13 +341,6 @@ main(int argc, char **argv) if (!opts->cparams.dbname && !opts->filename && !opts->tocSummary) pg_fatal("one of -d/--dbname and -f/--file must be specified"); - if (db_exclude_patterns.head != NULL && globals_only) - { - pg_log_error("option --exclude-database cannot be used together with -g/--globals-only"); - pg_log_error_hint("Try \"%s --help\" for more information.", progname); - exit_nicely(1); - } - /* Should get at most one of -d and -f, else user is confused */ if (opts->cparams.dbname) { @@ -417,13 +369,17 @@ main(int argc, char **argv) if (statistics_only && no_statistics) pg_fatal("options --statistics-only and --no-statistics cannot be used together"); - /* reject conflicting "with-" and "no-" options */ - if (with_data && no_data) - pg_fatal("options --with-data and --no-data cannot be used together"); - if (with_schema && no_schema) - pg_fatal("options --with-schema and --no-schema cannot be used together"); + /* reject conflicting "no-" options */ if (with_statistics && no_statistics) - pg_fatal("options --with-statistics and --no-statistics cannot be used together"); + pg_fatal("options --statistics and --no-statistics cannot be used together"); + + /* reject conflicting "only-" options */ + if (data_only && with_statistics) + pg_fatal("options %s and %s cannot be used together", + "-a/--data-only", "--statistics"); + if (schema_only && with_statistics) + pg_fatal("options %s and %s cannot be used together", + "-s/--schema-only", "--statistics"); if (data_only && opts->dropSchema) pg_fatal("options -c/--clean and -a/--data-only cannot be used together"); @@ -443,16 +399,14 @@ main(int argc, char **argv) pg_fatal("cannot specify both --single-transaction and multiple jobs"); /* - * Set derivative flags. An "-only" option may be overridden by an - * explicit "with-" option; e.g. "--schema-only --with-statistics" will - * include schema and statistics. Other ambiguous or nonsensical - * combinations, e.g. "--schema-only --no-schema", will have already - * caused an error in one of the checks above. + * Set derivative flags. Ambiguous or nonsensical combinations, e.g. + * "--schema-only --no-schema", will have already caused an error in one + * of the checks above. */ opts->dumpData = ((opts->dumpData && !schema_only && !statistics_only) || - (data_only || with_data)) && !no_data; + data_only) && !no_data; opts->dumpSchema = ((opts->dumpSchema && !data_only && !statistics_only) || - (schema_only || with_schema)) && !no_schema; + schema_only) && !no_schema; opts->dumpStatistics = ((opts->dumpStatistics && !schema_only && !data_only) || (statistics_only || with_statistics)) && !no_statistics; @@ -496,114 +450,6 @@ main(int argc, char **argv) opts->formatName); } - /* - * If toc.dat file is not present in the current path, then check for - * global.dat. If global.dat file is present, then restore all the - * databases from map.dat (if it exists), but skip restoring those - * matching --exclude-database patterns. - */ - if (inputFileSpec != NULL && !file_exists_in_directory(inputFileSpec, "toc.dat") && - file_exists_in_directory(inputFileSpec, "global.dat")) - { - PGconn *conn = NULL; /* Connection to restore global sql - * commands. */ - - /* - * Can only use --list or --use-list options with a single database - * dump. - */ - if (opts->tocSummary) - pg_fatal("option -l/--list cannot be used when restoring an archive created by pg_dumpall"); - else if (opts->tocFile) - pg_fatal("option -L/--use-list cannot be used when restoring an archive created by pg_dumpall"); - - /* - * To restore from a pg_dumpall archive, -C (create database) option - * must be specified unless we are only restoring globals. - */ - if (!globals_only && opts->createDB != 1) - { - pg_log_error("option -C/--create must be specified when restoring an archive created by pg_dumpall"); - pg_log_error_hint("Try \"%s --help\" for more information.", progname); - pg_log_error_hint("Individual databases can be restored using their specific archives."); - exit_nicely(1); - } - - /* - * Connect to the database to execute global sql commands from - * global.dat file. - */ - if (opts->cparams.dbname) - { - conn = ConnectDatabase(opts->cparams.dbname, NULL, opts->cparams.pghost, - opts->cparams.pgport, opts->cparams.username, TRI_DEFAULT, - false, progname, NULL, NULL, NULL, NULL); - - - if (!conn) - pg_fatal("could not connect to database \"%s\"", opts->cparams.dbname); - } - - /* If globals-only, then return from here. */ - if (globals_only) - { - /* - * Open global.dat file and execute/append all the global sql - * commands. - */ - n_errors = process_global_sql_commands(conn, inputFileSpec, - opts->filename); - - if (conn) - PQfinish(conn); - - pg_log_info("database restoring skipped because option -g/--globals-only was specified"); - } - else - { - /* Now restore all the databases from map.dat */ - n_errors = restore_all_databases(conn, inputFileSpec, db_exclude_patterns, - opts, numWorkers); - } - - /* Free db pattern list. */ - simple_string_list_destroy(&db_exclude_patterns); - } - else /* process if global.dat file does not exist. */ - { - if (db_exclude_patterns.head != NULL) - pg_fatal("option --exclude-database can be used only when restoring an archive created by pg_dumpall"); - - if (globals_only) - pg_fatal("option -g/--globals-only can be used only when restoring an archive created by pg_dumpall"); - - n_errors = restore_one_database(inputFileSpec, opts, numWorkers, false, 0); - } - - /* Done, print a summary of ignored errors during restore. */ - if (n_errors) - { - pg_log_warning("errors ignored on restore: %d", n_errors); - return 1; - } - - return 0; -} - -/* - * restore_one_database - * - * This will restore one database using toc.dat file. - * - * returns the number of errors while doing restore. - */ -static int -restore_one_database(const char *inputFileSpec, RestoreOptions *opts, - int numWorkers, bool append_data, int num) -{ - Archive *AH; - int n_errors; - AH = OpenArchive(inputFileSpec, opts->format); SetArchiveOptions(AH, NULL, opts); @@ -611,15 +457,9 @@ restore_one_database(const char *inputFileSpec, RestoreOptions *opts, /* * We don't have a connection yet but that doesn't matter. The connection * is initialized to NULL and if we terminate through exit_nicely() while - * it's still NULL, the cleanup function will just be a no-op. If we are - * restoring multiple databases, then only update AX handle for cleanup as - * the previous entry was already in the array and we had closed previous - * connection, so we can use the same array slot. + * it's still NULL, the cleanup function will just be a no-op. */ - if (!append_data || num == 0) - on_exit_close_archive(AH); - else - replace_on_exit_close_archive(AH); + on_exit_close_archive(AH); /* Let the archiver know how noisy to be */ AH->verbose = opts->verbose; @@ -639,21 +479,25 @@ restore_one_database(const char *inputFileSpec, RestoreOptions *opts, else { ProcessArchiveRestoreOptions(AH); - RestoreArchive(AH, append_data); + RestoreArchive(AH); } - n_errors = AH->n_errors; + /* done, print a summary of ignored errors */ + if (AH->n_errors) + pg_log_warning("errors ignored on restore: %d", AH->n_errors); /* AH may be freed in CloseArchive? */ + exit_code = AH->n_errors ? 1 : 0; + CloseArchive(AH); - return n_errors; + return exit_code; } static void usage(const char *progname) { - printf(_("%s restores PostgreSQL databases from archives created by pg_dump or pg_dumpall.\n\n"), progname); + printf(_("%s restores a PostgreSQL database from an archive created by pg_dump.\n\n"), progname); printf(_("Usage:\n")); printf(_(" %s [OPTION]... [FILE]\n"), progname); @@ -671,7 +515,6 @@ usage(const char *progname) printf(_(" -c, --clean clean (drop) database objects before recreating\n")); printf(_(" -C, --create create the target database\n")); printf(_(" -e, --exit-on-error exit on error, default is to continue\n")); - printf(_(" -g, --globals-only restore only global objects, no databases\n")); printf(_(" -I, --index=NAME restore named index\n")); printf(_(" -j, --jobs=NUM use this many parallel jobs to restore\n")); printf(_(" -L, --use-list=FILENAME use table of contents from this file for\n" @@ -688,7 +531,6 @@ usage(const char *progname) printf(_(" -1, --single-transaction restore as a single transaction\n")); printf(_(" --disable-triggers disable triggers during data-only restore\n")); printf(_(" --enable-row-security enable row security\n")); - printf(_(" --exclude-database=PATTERN do not restore the specified database(s)\n")); printf(_(" --filter=FILENAME restore or skip objects based on expressions\n" " in FILENAME\n")); printf(_(" --if-exists use IF EXISTS when dropping objects\n")); @@ -705,6 +547,7 @@ usage(const char *progname) printf(_(" --no-table-access-method do not restore table access methods\n")); printf(_(" --no-tablespaces do not restore tablespace assignments\n")); printf(_(" --section=SECTION restore named section (pre-data, data, or post-data)\n")); + printf(_(" --statistics restore the statistics\n")); printf(_(" --statistics-only restore only the statistics, not schema or data\n")); printf(_(" --strict-names require table and/or schema include patterns to\n" " match at least one entity each\n")); @@ -712,9 +555,6 @@ usage(const char *progname) printf(_(" --use-set-session-authorization\n" " use SET SESSION AUTHORIZATION commands instead of\n" " ALTER OWNER commands to set ownership\n")); - printf(_(" --with-data restore the data\n")); - printf(_(" --with-schema restore the schema\n")); - printf(_(" --with-statistics restore the statistics\n")); printf(_("\nConnection options:\n")); printf(_(" -h, --host=HOSTNAME database server host or socket directory\n")); @@ -725,8 +565,8 @@ usage(const char *progname) printf(_(" --role=ROLENAME do SET ROLE before restore\n")); printf(_("\n" - "The options -I, -n, -N, -P, -t, -T, --section, and --exclude-database can be\n" - "combined and specified multiple times to select multiple objects.\n")); + "The options -I, -n, -N, -P, -t, -T, and --section can be combined and specified\n" + "multiple times to select multiple objects.\n")); printf(_("\nIf no input file name is supplied, then standard input is used.\n\n")); printf(_("Report bugs to <%s>.\n"), PACKAGE_BUGREPORT); printf(_("%s home page: <%s>\n"), PACKAGE_NAME, PACKAGE_URL); @@ -831,585 +671,3 @@ read_restore_filters(const char *filename, RestoreOptions *opts) filter_free(&fstate); } - -/* - * file_exists_in_directory - * - * Returns true if the file exists in the given directory. - */ -static bool -file_exists_in_directory(const char *dir, const char *filename) -{ - struct stat st; - char buf[MAXPGPATH]; - - if (snprintf(buf, MAXPGPATH, "%s/%s", dir, filename) >= MAXPGPATH) - pg_fatal("directory name too long: \"%s\"", dir); - - return (stat(buf, &st) == 0 && S_ISREG(st.st_mode)); -} - -/* - * read_one_statement - * - * This will start reading from passed file pointer using fgetc and read till - * semicolon(sql statement terminator for global.dat file) - * - * EOF is returned if end-of-file input is seen; time to shut down. - */ - -static int -read_one_statement(StringInfo inBuf, FILE *pfile) -{ - int c; /* character read from getc() */ - int m; - - StringInfoData q; - - initStringInfo(&q); - - resetStringInfo(inBuf); - - /* - * Read characters until EOF or the appropriate delimiter is seen. - */ - while ((c = fgetc(pfile)) != EOF) - { - if (c != '\'' && c != '"' && c != '\n' && c != ';') - { - appendStringInfoChar(inBuf, (char) c); - while ((c = fgetc(pfile)) != EOF) - { - if (c != '\'' && c != '"' && c != ';' && c != '\n') - appendStringInfoChar(inBuf, (char) c); - else - break; - } - } - - if (c == '\'' || c == '"') - { - appendStringInfoChar(&q, (char) c); - m = c; - - while ((c = fgetc(pfile)) != EOF) - { - appendStringInfoChar(&q, (char) c); - - if (c == m) - { - appendStringInfoString(inBuf, q.data); - resetStringInfo(&q); - break; - } - } - } - - if (c == ';') - { - appendStringInfoChar(inBuf, (char) ';'); - break; - } - - if (c == '\n') - appendStringInfoChar(inBuf, (char) '\n'); - } - - pg_free(q.data); - - /* No input before EOF signal means time to quit. */ - if (c == EOF && inBuf->len == 0) - return EOF; - - /* return something that's not EOF */ - return 'Q'; -} - -/* - * get_dbnames_list_to_restore - * - * This will mark for skipping any entries from dbname_oid_list that pattern match an - * entry in the db_exclude_patterns list. - * - * Returns the number of database to be restored. - * - */ -static int -get_dbnames_list_to_restore(PGconn *conn, - SimplePtrList *dbname_oid_list, - SimpleStringList db_exclude_patterns) -{ - int count_db = 0; - PQExpBuffer query; - PGresult *res; - - query = createPQExpBuffer(); - - if (!conn) - pg_log_info("considering PATTERN as NAME for --exclude-database option as no database connection while doing pg_restore"); - - /* - * Process one by one all dbnames and if specified to skip restoring, then - * remove dbname from list. - */ - for (SimplePtrListCell *db_cell = dbname_oid_list->head; - db_cell; db_cell = db_cell->next) - { - DbOidName *dbidname = (DbOidName *) db_cell->ptr; - bool skip_db_restore = false; - PQExpBuffer db_lit = createPQExpBuffer(); - - appendStringLiteralConn(db_lit, dbidname->str, conn); - - for (SimpleStringListCell *pat_cell = db_exclude_patterns.head; pat_cell; pat_cell = pat_cell->next) - { - /* - * If there is an exact match then we don't need to try a pattern - * match - */ - if (pg_strcasecmp(dbidname->str, pat_cell->val) == 0) - skip_db_restore = true; - /* Otherwise, try a pattern match if there is a connection */ - else if (conn) - { - int dotcnt; - - appendPQExpBufferStr(query, "SELECT 1 "); - processSQLNamePattern(conn, query, pat_cell->val, false, - false, NULL, db_lit->data, - NULL, NULL, NULL, &dotcnt); - - if (dotcnt > 0) - { - pg_log_error("improper qualified name (too many dotted names): %s", - dbidname->str); - PQfinish(conn); - exit_nicely(1); - } - - res = executeQuery(conn, query->data); - - if ((PQresultStatus(res) == PGRES_TUPLES_OK) && PQntuples(res)) - { - skip_db_restore = true; - pg_log_info("database name \"%s\" matches exclude pattern \"%s\"", dbidname->str, pat_cell->val); - } - - PQclear(res); - resetPQExpBuffer(query); - } - - if (skip_db_restore) - break; - } - - destroyPQExpBuffer(db_lit); - - /* - * Mark db to be skipped or increment the counter of dbs to be - * restored - */ - if (skip_db_restore) - { - pg_log_info("excluding database \"%s\"", dbidname->str); - dbidname->oid = InvalidOid; - } - else - { - count_db++; - } - } - - destroyPQExpBuffer(query); - - return count_db; -} - -/* - * get_dbname_oid_list_from_mfile - * - * Open map.dat file and read line by line and then prepare a list of database - * names and corresponding db_oid. - * - * Returns, total number of database names in map.dat file. - */ -static int -get_dbname_oid_list_from_mfile(const char *dumpdirpath, SimplePtrList *dbname_oid_list) -{ - StringInfoData linebuf; - FILE *pfile; - char map_file_path[MAXPGPATH]; - int count = 0; - - - /* - * If there is only global.dat file in dump, then return from here as - * there is no database to restore. - */ - if (!file_exists_in_directory(dumpdirpath, "map.dat")) - { - pg_log_info("database restoring is skipped because file \"%s\" does not exist in directory \"%s\"", "map.dat", dumpdirpath); - return 0; - } - - snprintf(map_file_path, MAXPGPATH, "%s/map.dat", dumpdirpath); - - /* Open map.dat file. */ - pfile = fopen(map_file_path, PG_BINARY_R); - - if (pfile == NULL) - pg_fatal("could not open file \"%s\": %m", map_file_path); - - initStringInfo(&linebuf); - - /* Append all the dbname/db_oid combinations to the list. */ - while (pg_get_line_buf(pfile, &linebuf)) - { - Oid db_oid = InvalidOid; - char *dbname; - DbOidName *dbidname; - int namelen; - char *p = linebuf.data; - - /* Extract dboid. */ - while (isdigit((unsigned char) *p)) - p++; - if (p > linebuf.data && *p == ' ') - { - sscanf(linebuf.data, "%u", &db_oid); - p++; - } - - /* dbname is the rest of the line */ - dbname = p; - namelen = strlen(dbname); - - /* Report error and exit if the file has any corrupted data. */ - if (!OidIsValid(db_oid) || namelen <= 1) - pg_fatal("invalid entry in file \"%s\" on line %d", map_file_path, - count + 1); - - pg_log_info("found database \"%s\" (OID: %u) in file \"%s\"", - dbname, db_oid, map_file_path); - - dbidname = pg_malloc(offsetof(DbOidName, str) + namelen + 1); - dbidname->oid = db_oid; - strlcpy(dbidname->str, dbname, namelen); - - simple_ptr_list_append(dbname_oid_list, dbidname); - count++; - } - - /* Close map.dat file. */ - fclose(pfile); - - return count; -} - -/* - * restore_all_databases - * - * This will restore databases those dumps are present in - * directory based on map.dat file mapping. - * - * This will skip restoring for databases that are specified with - * exclude-database option. - * - * returns, number of errors while doing restore. - */ -static int -restore_all_databases(PGconn *conn, const char *dumpdirpath, - SimpleStringList db_exclude_patterns, RestoreOptions *opts, - int numWorkers) -{ - SimplePtrList dbname_oid_list = {NULL, NULL}; - int num_db_restore = 0; - int num_total_db; - int n_errors_total; - int count = 0; - char *connected_db = NULL; - bool dumpData = opts->dumpData; - bool dumpSchema = opts->dumpSchema; - bool dumpStatistics = opts->dumpSchema; - - /* Save db name to reuse it for all the database. */ - if (opts->cparams.dbname) - connected_db = opts->cparams.dbname; - - num_total_db = get_dbname_oid_list_from_mfile(dumpdirpath, &dbname_oid_list); - - /* If map.dat has no entries, return after processing global.dat */ - if (dbname_oid_list.head == NULL) - return process_global_sql_commands(conn, dumpdirpath, opts->filename); - - pg_log_info(ngettext("found %d database name in \"%s\"", - "found %d database names in \"%s\"", - num_total_db), - num_total_db, "map.dat"); - - if (!conn) - { - pg_log_info("trying to connect to database \"%s\"", "postgres"); - - conn = ConnectDatabase("postgres", NULL, opts->cparams.pghost, - opts->cparams.pgport, opts->cparams.username, TRI_DEFAULT, - false, progname, NULL, NULL, NULL, NULL); - - /* Try with template1. */ - if (!conn) - { - pg_log_info("trying to connect to database \"%s\"", "template1"); - - conn = ConnectDatabase("template1", NULL, opts->cparams.pghost, - opts->cparams.pgport, opts->cparams.username, TRI_DEFAULT, - false, progname, NULL, NULL, NULL, NULL); - } - } - - /* - * filter the db list according to the exclude patterns - */ - num_db_restore = get_dbnames_list_to_restore(conn, &dbname_oid_list, - db_exclude_patterns); - - /* Open global.dat file and execute/append all the global sql commands. */ - n_errors_total = process_global_sql_commands(conn, dumpdirpath, opts->filename); - - /* Close the db connection as we are done with globals and patterns. */ - if (conn) - PQfinish(conn); - - /* Exit if no db needs to be restored. */ - if (dbname_oid_list.head == NULL || num_db_restore == 0) - { - pg_log_info(ngettext("no database needs restoring out of %d database", - "no database needs restoring out of %d databases", num_total_db), - num_total_db); - return n_errors_total; - } - - pg_log_info("need to restore %d databases out of %d databases", num_db_restore, num_total_db); - - /* - * We have a list of databases to restore after processing the - * exclude-database switch(es). Now we can restore them one by one. - */ - for (SimplePtrListCell *db_cell = dbname_oid_list.head; - db_cell; db_cell = db_cell->next) - { - DbOidName *dbidname = (DbOidName *) db_cell->ptr; - char subdirpath[MAXPGPATH]; - char subdirdbpath[MAXPGPATH]; - char dbfilename[MAXPGPATH]; - int n_errors; - - /* ignore dbs marked for skipping */ - if (dbidname->oid == InvalidOid) - continue; - - /* - * We need to reset override_dbname so that objects can be restored - * into an already created database. (used with -d/--dbname option) - */ - if (opts->cparams.override_dbname) - { - pfree(opts->cparams.override_dbname); - opts->cparams.override_dbname = NULL; - } - - snprintf(subdirdbpath, MAXPGPATH, "%s/databases", dumpdirpath); - - /* - * Look for the database dump file/dir. If there is an {oid}.tar or - * {oid}.dmp file, use it. Otherwise try to use a directory called - * {oid} - */ - snprintf(dbfilename, MAXPGPATH, "%u.tar", dbidname->oid); - if (file_exists_in_directory(subdirdbpath, dbfilename)) - snprintf(subdirpath, MAXPGPATH, "%s/databases/%u.tar", dumpdirpath, dbidname->oid); - else - { - snprintf(dbfilename, MAXPGPATH, "%u.dmp", dbidname->oid); - - if (file_exists_in_directory(subdirdbpath, dbfilename)) - snprintf(subdirpath, MAXPGPATH, "%s/databases/%u.dmp", dumpdirpath, dbidname->oid); - else - snprintf(subdirpath, MAXPGPATH, "%s/databases/%u", dumpdirpath, dbidname->oid); - } - - pg_log_info("restoring database \"%s\"", dbidname->str); - - /* If database is already created, then don't set createDB flag. */ - if (opts->cparams.dbname) - { - PGconn *test_conn; - - test_conn = ConnectDatabase(dbidname->str, NULL, opts->cparams.pghost, - opts->cparams.pgport, opts->cparams.username, TRI_DEFAULT, - false, progname, NULL, NULL, NULL, NULL); - if (test_conn) - { - PQfinish(test_conn); - - /* Use already created database for connection. */ - opts->createDB = 0; - opts->cparams.dbname = dbidname->str; - } - else - { - /* we'll have to create it */ - opts->createDB = 1; - opts->cparams.dbname = connected_db; - } - } - - /* - * Reset flags - might have been reset in pg_backup_archiver.c by the - * previous restore. - */ - opts->dumpData = dumpData; - opts->dumpSchema = dumpSchema; - opts->dumpStatistics = dumpStatistics; - - /* Restore the single database. */ - n_errors = restore_one_database(subdirpath, opts, numWorkers, true, count); - - /* Print a summary of ignored errors during single database restore. */ - if (n_errors) - { - n_errors_total += n_errors; - pg_log_warning("errors ignored on database \"%s\" restore: %d", dbidname->str, n_errors); - } - - count++; - } - - /* Log number of processed databases. */ - pg_log_info("number of restored databases is %d", num_db_restore); - - /* Free dbname and dboid list. */ - simple_ptr_list_destroy(&dbname_oid_list); - - return n_errors_total; -} - -/* - * process_global_sql_commands - * - * Open global.dat and execute or copy the sql commands one by one. - * - * If outfile is not NULL, copy all sql commands into outfile rather than - * executing them. - * - * Returns the number of errors while processing global.dat - */ -static int -process_global_sql_commands(PGconn *conn, const char *dumpdirpath, const char *outfile) -{ - char global_file_path[MAXPGPATH]; - PGresult *result; - StringInfoData sqlstatement, - user_create; - FILE *pfile; - int n_errors = 0; - - snprintf(global_file_path, MAXPGPATH, "%s/global.dat", dumpdirpath); - - /* Open global.dat file. */ - pfile = fopen(global_file_path, PG_BINARY_R); - - if (pfile == NULL) - pg_fatal("could not open file \"%s\": %m", global_file_path); - - /* - * If outfile is given, then just copy all global.dat file data into - * outfile. - */ - if (outfile) - { - copy_or_print_global_file(outfile, pfile); - return 0; - } - - /* Init sqlstatement to append commands. */ - initStringInfo(&sqlstatement); - - /* creation statement for our current role */ - initStringInfo(&user_create); - appendStringInfoString(&user_create, "CREATE ROLE "); - /* should use fmtId here, but we don't know the encoding */ - appendStringInfoString(&user_create, PQuser(conn)); - appendStringInfoChar(&user_create, ';'); - - /* Process file till EOF and execute sql statements. */ - while (read_one_statement(&sqlstatement, pfile) != EOF) - { - /* don't try to create the role we are connected as */ - if (strstr(sqlstatement.data, user_create.data)) - continue; - - pg_log_info("executing query: %s", sqlstatement.data); - result = PQexec(conn, sqlstatement.data); - - switch (PQresultStatus(result)) - { - case PGRES_COMMAND_OK: - case PGRES_TUPLES_OK: - case PGRES_EMPTY_QUERY: - break; - default: - n_errors++; - pg_log_error("could not execute query: %s", PQerrorMessage(conn)); - pg_log_error_detail("Command was: %s", sqlstatement.data); - } - PQclear(result); - } - - /* Print a summary of ignored errors during global.dat. */ - if (n_errors) - pg_log_warning(ngettext("ignored %d error in file \"%s\"", - "ignored %d errors in file \"%s\"", n_errors), - n_errors, global_file_path); - fclose(pfile); - - return n_errors; -} - -/* - * copy_or_print_global_file - * - * Copy global.dat into the output file. If "-" is used as outfile, - * then print commands to stdout. - */ -static void -copy_or_print_global_file(const char *outfile, FILE *pfile) -{ - char out_file_path[MAXPGPATH]; - FILE *OPF; - int c; - - /* "-" is used for stdout. */ - if (strcmp(outfile, "-") == 0) - OPF = stdout; - else - { - snprintf(out_file_path, MAXPGPATH, "%s", outfile); - OPF = fopen(out_file_path, PG_BINARY_W); - - if (OPF == NULL) - { - fclose(pfile); - pg_fatal("could not open file: \"%s\"", outfile); - } - } - - /* Append global.dat into output file or print to stdout. */ - while ((c = fgetc(pfile)) != EOF) - fputc(c, OPF); - - fclose(pfile); - - /* Close output file. */ - if (strcmp(outfile, "-") != 0) - fclose(OPF); -} diff --git a/src/bin/pg_dump/t/001_basic.pl b/src/bin/pg_dump/t/001_basic.pl index c3c5fae11eaaf..37d893d5e6a5f 100644 --- a/src/bin/pg_dump/t/001_basic.pl +++ b/src/bin/pg_dump/t/001_basic.pl @@ -237,24 +237,6 @@ 'pg_restore: options -C\/--create and -1\/--single-transaction cannot be used together' ); -command_fails_like( - [ 'pg_restore', '--exclude-database=foo', '--globals-only', '-d', 'xxx' ], - qr/\Qpg_restore: error: option --exclude-database cannot be used together with -g\/--globals-only\E/, - 'pg_restore: option --exclude-database cannot be used together with -g/--globals-only' -); - -command_fails_like( - [ 'pg_restore', '--exclude-database=foo', '-d', 'xxx', 'dumpdir' ], - qr/\Qpg_restore: error: option --exclude-database can be used only when restoring an archive created by pg_dumpall\E/, - 'When option --exclude-database is used in pg_restore with dump of pg_dump' -); - -command_fails_like( - [ 'pg_restore', '--globals-only', '-d', 'xxx', 'dumpdir' ], - qr/\Qpg_restore: error: option -g\/--globals-only can be used only when restoring an archive created by pg_dumpall\E/, - 'When option --globals-only is not used in pg_restore with dump of pg_dump' -); - # also fails for -r and -t, but it seems pointless to add more tests for those. command_fails_like( [ 'pg_dumpall', '--exclude-database=foo', '--globals-only' ], @@ -262,8 +244,4 @@ 'pg_dumpall: option --exclude-database cannot be used together with -g/--globals-only' ); -command_fails_like( - [ 'pg_dumpall', '--format', 'x' ], - qr/\Qpg_dumpall: error: unrecognized output format "x";\E/, - 'pg_dumpall: unrecognized output format'); done_testing(); diff --git a/src/bin/pg_dump/t/002_pg_dump.pl b/src/bin/pg_dump/t/002_pg_dump.pl index 6c7ec80e271ce..a86b38466de14 100644 --- a/src/bin/pg_dump/t/002_pg_dump.pl +++ b/src/bin/pg_dump/t/002_pg_dump.pl @@ -68,7 +68,7 @@ '--no-data', '--sequence-data', '--binary-upgrade', - '--with-statistics', + '--statistics', '--dbname' => 'postgres', # alternative way to specify database ], restore_cmd => [ @@ -76,7 +76,7 @@ '--format' => 'custom', '--verbose', '--file' => "$tempdir/binary_upgrade.sql", - '--with-statistics', + '--statistics', "$tempdir/binary_upgrade.dump", ], }, @@ -90,13 +90,13 @@ '--format' => 'custom', '--compress' => '1', '--file' => "$tempdir/compression_gzip_custom.dump", - '--with-statistics', + '--statistics', 'postgres', ], restore_cmd => [ 'pg_restore', '--file' => "$tempdir/compression_gzip_custom.sql", - '--with-statistics', + '--statistics', "$tempdir/compression_gzip_custom.dump", ], command_like => { @@ -119,7 +119,7 @@ '--format' => 'directory', '--compress' => 'gzip:1', '--file' => "$tempdir/compression_gzip_dir", - '--with-statistics', + '--statistics', 'postgres', ], # Give coverage for manually compressed blobs.toc files during @@ -137,7 +137,7 @@ 'pg_restore', '--jobs' => '2', '--file' => "$tempdir/compression_gzip_dir.sql", - '--with-statistics', + '--statistics', "$tempdir/compression_gzip_dir", ], }, @@ -150,7 +150,7 @@ '--format' => 'plain', '--compress' => '1', '--file' => "$tempdir/compression_gzip_plain.sql.gz", - '--with-statistics', + '--statistics', 'postgres', ], # Decompress the generated file to run through the tests. @@ -169,13 +169,13 @@ '--format' => 'custom', '--compress' => 'lz4', '--file' => "$tempdir/compression_lz4_custom.dump", - '--with-statistics', + '--statistics', 'postgres', ], restore_cmd => [ 'pg_restore', '--file' => "$tempdir/compression_lz4_custom.sql", - '--with-statistics', + '--statistics', "$tempdir/compression_lz4_custom.dump", ], command_like => { @@ -198,7 +198,7 @@ '--format' => 'directory', '--compress' => 'lz4:1', '--file' => "$tempdir/compression_lz4_dir", - '--with-statistics', + '--statistics', 'postgres', ], # Verify that data files were compressed @@ -210,7 +210,7 @@ 'pg_restore', '--jobs' => '2', '--file' => "$tempdir/compression_lz4_dir.sql", - '--with-statistics', + '--statistics', "$tempdir/compression_lz4_dir", ], }, @@ -223,7 +223,7 @@ '--format' => 'plain', '--compress' => 'lz4', '--file' => "$tempdir/compression_lz4_plain.sql.lz4", - '--with-statistics', + '--statistics', 'postgres', ], # Decompress the generated file to run through the tests. @@ -245,13 +245,13 @@ '--format' => 'custom', '--compress' => 'zstd', '--file' => "$tempdir/compression_zstd_custom.dump", - '--with-statistics', + '--statistics', 'postgres', ], restore_cmd => [ 'pg_restore', '--file' => "$tempdir/compression_zstd_custom.sql", - '--with-statistics', + '--statistics', "$tempdir/compression_zstd_custom.dump", ], command_like => { @@ -273,7 +273,7 @@ '--format' => 'directory', '--compress' => 'zstd:1', '--file' => "$tempdir/compression_zstd_dir", - '--with-statistics', + '--statistics', 'postgres', ], # Give coverage for manually compressed blobs.toc files during @@ -294,7 +294,7 @@ 'pg_restore', '--jobs' => '2', '--file' => "$tempdir/compression_zstd_dir.sql", - '--with-statistics', + '--statistics', "$tempdir/compression_zstd_dir", ], }, @@ -308,7 +308,7 @@ '--format' => 'plain', '--compress' => 'zstd:long', '--file' => "$tempdir/compression_zstd_plain.sql.zst", - '--with-statistics', + '--statistics', 'postgres', ], # Decompress the generated file to run through the tests. @@ -327,7 +327,7 @@ 'pg_dump', '--no-sync', '--file' => "$tempdir/clean.sql", '--clean', - '--with-statistics', + '--statistics', '--dbname' => 'postgres', # alternative way to specify database ], }, @@ -338,7 +338,7 @@ '--clean', '--if-exists', '--encoding' => 'UTF8', # no-op, just for testing - '--with-statistics', + '--statistics', 'postgres', ], }, @@ -357,7 +357,7 @@ '--create', '--no-reconnect', # no-op, just for testing '--verbose', - '--with-statistics', + '--statistics', 'postgres', ], }, @@ -376,7 +376,7 @@ dump_cmd => [ 'pg_dump', '--no-sync', '--file' => "$tempdir/defaults.sql", - '--with-statistics', + '--statistics', 'postgres', ], }, @@ -385,7 +385,7 @@ dump_cmd => [ 'pg_dump', '--no-sync', '--file' => "$tempdir/defaults_no_public.sql", - '--with-statistics', + '--statistics', 'regress_pg_dump_test', ], }, @@ -395,7 +395,7 @@ 'pg_dump', '--no-sync', '--clean', '--file' => "$tempdir/defaults_no_public_clean.sql", - '--with-statistics', + '--statistics', 'regress_pg_dump_test', ], }, @@ -404,7 +404,7 @@ dump_cmd => [ 'pg_dump', '--no-sync', '--file' => "$tempdir/defaults_public_owner.sql", - '--with-statistics', + '--statistics', 'regress_public_owner', ], }, @@ -419,14 +419,14 @@ 'pg_dump', '--format' => 'custom', '--file' => "$tempdir/defaults_custom_format.dump", - '--with-statistics', + '--statistics', 'postgres', ], restore_cmd => [ 'pg_restore', '--format' => 'custom', '--file' => "$tempdir/defaults_custom_format.sql", - '--with-statistics', + '--statistics', "$tempdir/defaults_custom_format.dump", ], command_like => { @@ -451,14 +451,14 @@ 'pg_dump', '--format' => 'directory', '--file' => "$tempdir/defaults_dir_format", - '--with-statistics', + '--statistics', 'postgres', ], restore_cmd => [ 'pg_restore', '--format' => 'directory', '--file' => "$tempdir/defaults_dir_format.sql", - '--with-statistics', + '--statistics', "$tempdir/defaults_dir_format", ], command_like => { @@ -484,13 +484,13 @@ '--format' => 'directory', '--jobs' => 2, '--file' => "$tempdir/defaults_parallel", - '--with-statistics', + '--statistics', 'postgres', ], restore_cmd => [ 'pg_restore', '--file' => "$tempdir/defaults_parallel.sql", - '--with-statistics', + '--statistics', "$tempdir/defaults_parallel", ], }, @@ -502,14 +502,14 @@ 'pg_dump', '--format' => 'tar', '--file' => "$tempdir/defaults_tar_format.tar", - '--with-statistics', + '--statistics', 'postgres', ], restore_cmd => [ 'pg_restore', '--format' => 'tar', '--file' => "$tempdir/defaults_tar_format.sql", - '--with-statistics', + '--statistics', "$tempdir/defaults_tar_format.tar", ], }, @@ -518,7 +518,7 @@ 'pg_dump', '--no-sync', '--file' => "$tempdir/exclude_dump_test_schema.sql", '--exclude-schema' => 'dump_test', - '--with-statistics', + '--statistics', 'postgres', ], }, @@ -527,7 +527,7 @@ 'pg_dump', '--no-sync', '--file' => "$tempdir/exclude_test_table.sql", '--exclude-table' => 'dump_test.test_table', - '--with-statistics', + '--statistics', 'postgres', ], }, @@ -536,7 +536,7 @@ 'pg_dump', '--no-sync', '--file' => "$tempdir/exclude_measurement.sql", '--exclude-table-and-children' => 'dump_test.measurement', - '--with-statistics', + '--statistics', 'postgres', ], }, @@ -546,7 +546,7 @@ '--file' => "$tempdir/exclude_measurement_data.sql", '--exclude-table-data-and-children' => 'dump_test.measurement', '--no-unlogged-table-data', - '--with-statistics', + '--statistics', 'postgres', ], }, @@ -556,7 +556,7 @@ '--file' => "$tempdir/exclude_test_table_data.sql", '--exclude-table-data' => 'dump_test.test_table', '--no-unlogged-table-data', - '--with-statistics', + '--statistics', 'postgres', ], }, @@ -575,7 +575,7 @@ '--file' => "$tempdir/pg_dumpall_globals.sql", '--globals-only', '--no-sync', - '--with-statistics', + '--statistics', ], }, pg_dumpall_globals_clean => { @@ -585,14 +585,14 @@ '--globals-only', '--clean', '--no-sync', - '--with-statistics', + '--statistics', ], }, pg_dumpall_dbprivs => { dump_cmd => [ 'pg_dumpall', '--no-sync', '--file' => "$tempdir/pg_dumpall_dbprivs.sql", - '--with-statistics', + '--statistics', ], }, pg_dumpall_exclude => { @@ -602,7 +602,7 @@ '--file' => "$tempdir/pg_dumpall_exclude.sql", '--exclude-database' => '*dump_test*', '--no-sync', - '--with-statistics', + '--statistics', ], }, no_toast_compression => { @@ -610,7 +610,7 @@ 'pg_dump', '--no-sync', '--file' => "$tempdir/no_toast_compression.sql", '--no-toast-compression', - '--with-statistics', + '--statistics', 'postgres', ], }, @@ -619,7 +619,7 @@ 'pg_dump', '--no-sync', '--file' => "$tempdir/no_large_objects.sql", '--no-large-objects', - '--with-statistics', + '--statistics', 'postgres', ], }, @@ -628,7 +628,7 @@ 'pg_dump', '--no-sync', '--file' => "$tempdir/no_policies.sql", '--no-policies', - '--with-statistics', + '--statistics', 'postgres', ], }, @@ -637,7 +637,7 @@ 'pg_dump', '--no-sync', '--file' => "$tempdir/no_privs.sql", '--no-privileges', - '--with-statistics', + '--statistics', 'postgres', ], }, @@ -646,7 +646,7 @@ 'pg_dump', '--no-sync', '--file' => "$tempdir/no_owner.sql", '--no-owner', - '--with-statistics', + '--statistics', 'postgres', ], }, @@ -655,7 +655,7 @@ 'pg_dump', '--no-sync', '--file' => "$tempdir/no_table_access_method.sql", '--no-table-access-method', - '--with-statistics', + '--statistics', 'postgres', ], }, @@ -664,7 +664,7 @@ 'pg_dump', '--no-sync', '--file' => "$tempdir/only_dump_test_schema.sql", '--schema' => 'dump_test', - '--with-statistics', + '--statistics', 'postgres', ], }, @@ -675,7 +675,7 @@ '--table' => 'dump_test.test_table', '--lock-wait-timeout' => (1000 * $PostgreSQL::Test::Utils::timeout_default), - '--with-statistics', + '--statistics', 'postgres', ], }, @@ -686,7 +686,7 @@ '--table-and-children' => 'dump_test.measurement', '--lock-wait-timeout' => (1000 * $PostgreSQL::Test::Utils::timeout_default), - '--with-statistics', + '--statistics', 'postgres', ], }, @@ -696,7 +696,7 @@ '--file' => "$tempdir/role.sql", '--role' => 'regress_dump_test_role', '--schema' => 'dump_test_second_schema', - '--with-statistics', + '--statistics', 'postgres', ], }, @@ -709,13 +709,13 @@ '--file' => "$tempdir/role_parallel", '--role' => 'regress_dump_test_role', '--schema' => 'dump_test_second_schema', - '--with-statistics', + '--statistics', 'postgres', ], restore_cmd => [ 'pg_restore', '--file' => "$tempdir/role_parallel.sql", - '--with-statistics', + '--statistics', "$tempdir/role_parallel", ], }, @@ -744,7 +744,7 @@ 'pg_dump', '--no-sync', '--file' => "$tempdir/section_pre_data.sql", '--section' => 'pre-data', - '--with-statistics', + '--statistics', 'postgres', ], }, @@ -753,7 +753,7 @@ 'pg_dump', '--no-sync', '--file' => "$tempdir/section_data.sql", '--section' => 'data', - '--with-statistics', + '--statistics', 'postgres', ], }, @@ -762,7 +762,7 @@ 'pg_dump', '--no-sync', '--file' => "$tempdir/section_post_data.sql", '--section' => 'post-data', - '--with-statistics', + '--statistics', 'postgres', ], }, @@ -773,7 +773,7 @@ '--schema' => 'dump_test', '--large-objects', '--no-large-objects', - '--with-statistics', + '--statistics', 'postgres', ], }, @@ -789,7 +789,7 @@ 'pg_dump', '--no-sync', "--file=$tempdir/no_data_no_schema.sql", '--no-data', '--no-schema', 'postgres', - '--with-statistics', + '--statistics', ], }, statistics_only => { @@ -799,18 +799,11 @@ 'postgres', ], }, - schema_only_with_statistics => { - dump_cmd => [ - 'pg_dump', '--no-sync', - "--file=$tempdir/schema_only_with_statistics.sql", - '--schema-only', '--with-statistics', 'postgres', - ], - }, no_schema => { dump_cmd => [ 'pg_dump', '--no-sync', "--file=$tempdir/no_schema.sql", '--no-schema', - '--with-statistics', 'postgres', + '--statistics', 'postgres', ], },); @@ -5207,6 +5200,17 @@ qr/\Qpg_dump: error: no matching schemas were found for pattern\E/, 'no matching schemas'); +command_fails_like( + [ + 'pg_dump', + '--port' => $port, + '--strict-names', + '--schema-only', + '--statistics', + ], + qr/\Qpg_dump: error: options -s\/--schema-only and --statistics cannot be used together\E/, + 'cannot use --schema-only and --statistics together'); + command_fails_like( [ 'pg_dump', diff --git a/src/bin/pg_dump/t/006_pg_dumpall.pl b/src/bin/pg_dump/t/006_pg_dumpall.pl deleted file mode 100644 index c274b777586ad..0000000000000 --- a/src/bin/pg_dump/t/006_pg_dumpall.pl +++ /dev/null @@ -1,400 +0,0 @@ -# Copyright (c) 2021-2025, PostgreSQL Global Development Group - -use strict; -use warnings FATAL => 'all'; - -use PostgreSQL::Test::Cluster; -use PostgreSQL::Test::Utils; -use Test::More; - -my $tempdir = PostgreSQL::Test::Utils::tempdir; -my $run_db = 'postgres'; -my $sep = $windows_os ? "\\" : "/"; - -# Tablespace locations used by "restore_tablespace" test case. -my $tablespace1 = "${tempdir}${sep}tbl1"; -my $tablespace2 = "${tempdir}${sep}tbl2"; -mkdir($tablespace1) || die "mkdir $tablespace1 $!"; -mkdir($tablespace2) || die "mkdir $tablespace2 $!"; - -# Scape tablespace locations on Windows. -$tablespace1 = $windows_os ? ($tablespace1 =~ s/\\/\\\\/gr) : $tablespace1; -$tablespace2 = $windows_os ? ($tablespace2 =~ s/\\/\\\\/gr) : $tablespace2; - -# Where pg_dumpall will be executed. -my $node = PostgreSQL::Test::Cluster->new('node'); -$node->init; -$node->start; - - -############################################################### -# Definition of the pg_dumpall test cases to run. -# -# Each of these test cases are named and those names are used for fail -# reporting and also to save the dump and restore information needed for the -# test to assert. -# -# The "setup_sql" is a psql valid script that contains SQL commands to execute -# before of actually execute the tests. The setups are all executed before of -# any test execution. -# -# The "dump_cmd" and "restore_cmd" are the commands that will be executed. The -# "restore_cmd" must have the --file flag to save the restore output so that we -# can assert on it. -# -# The "like" and "unlike" is a regexp that is used to match the pg_restore -# output. It must have at least one of then filled per test cases but it also -# can have both. See "excluding_databases" test case for example. -my %pgdumpall_runs = ( - restore_roles => { - setup_sql => ' - CREATE ROLE dumpall WITH ENCRYPTED PASSWORD \'admin\' SUPERUSER; - CREATE ROLE dumpall2 WITH REPLICATION CONNECTION LIMIT 10;', - dump_cmd => [ - 'pg_dumpall', - '--format' => 'directory', - '--file' => "$tempdir/restore_roles", - ], - restore_cmd => [ - 'pg_restore', '-C', - '--format' => 'directory', - '--file' => "$tempdir/restore_roles.sql", - "$tempdir/restore_roles", - ], - like => qr/ - ^\s*\QCREATE ROLE dumpall;\E\s*\n - \s*\QALTER ROLE dumpall WITH SUPERUSER INHERIT NOCREATEROLE NOCREATEDB NOLOGIN NOREPLICATION NOBYPASSRLS PASSWORD 'SCRAM-SHA-256\E - [^']+';\s*\n - \s*\QCREATE ROLE dumpall2;\E - \s*\QALTER ROLE dumpall2 WITH NOSUPERUSER INHERIT NOCREATEROLE NOCREATEDB NOLOGIN REPLICATION NOBYPASSRLS CONNECTION LIMIT 10;\E - /xm - }, - - restore_tablespace => { - setup_sql => " - CREATE ROLE tap; - CREATE TABLESPACE tbl1 OWNER tap LOCATION '$tablespace1'; - CREATE TABLESPACE tbl2 OWNER tap LOCATION '$tablespace2' WITH (seq_page_cost=1.0);", - dump_cmd => [ - 'pg_dumpall', - '--format' => 'directory', - '--file' => "$tempdir/restore_tablespace", - ], - restore_cmd => [ - 'pg_restore', '-C', - '--format' => 'directory', - '--file' => "$tempdir/restore_tablespace.sql", - "$tempdir/restore_tablespace", - ], - # Match "E" as optional since it is added on LOCATION when running on - # Windows. - like => qr/^ - \n\QCREATE TABLESPACE tbl1 OWNER tap LOCATION \E(?:E)?\Q'$tablespace1';\E - \n\QCREATE TABLESPACE tbl2 OWNER tap LOCATION \E(?:E)?\Q'$tablespace2';\E - \n\QALTER TABLESPACE tbl2 SET (seq_page_cost=1.0);\E - /xm, - }, - - restore_grants => { - setup_sql => " - CREATE DATABASE tapgrantsdb; - CREATE SCHEMA private; - CREATE SEQUENCE serial START 101; - CREATE FUNCTION fn() RETURNS void AS \$\$ - BEGIN - END; - \$\$ LANGUAGE plpgsql; - CREATE ROLE super; - CREATE ROLE grant1; - CREATE ROLE grant2; - CREATE ROLE grant3; - CREATE ROLE grant4; - CREATE ROLE grant5; - CREATE ROLE grant6; - CREATE ROLE grant7; - CREATE ROLE grant8; - - CREATE TABLE t (id int); - INSERT INTO t VALUES (1), (2), (3), (4); - - GRANT SELECT ON TABLE t TO grant1; - GRANT INSERT ON TABLE t TO grant2; - GRANT ALL PRIVILEGES ON TABLE t to grant3; - GRANT CONNECT, CREATE ON DATABASE tapgrantsdb TO grant4; - GRANT USAGE, CREATE ON SCHEMA private TO grant5; - GRANT USAGE, SELECT, UPDATE ON SEQUENCE serial TO grant6; - GRANT super TO grant7; - GRANT EXECUTE ON FUNCTION fn() TO grant8; - ", - dump_cmd => [ - 'pg_dumpall', - '--format' => 'directory', - '--file' => "$tempdir/restore_grants", - ], - restore_cmd => [ - 'pg_restore', '-C', - '--format' => 'directory', - '--file' => "$tempdir/restore_grants.sql", - "$tempdir/restore_grants", - ], - like => qr/^ - \n\QGRANT super TO grant7 WITH INHERIT TRUE GRANTED BY\E - (.*\n)* - \n\QGRANT ALL ON SCHEMA private TO grant5;\E - (.*\n)* - \n\QGRANT ALL ON FUNCTION public.fn() TO grant8;\E - (.*\n)* - \n\QGRANT ALL ON SEQUENCE public.serial TO grant6;\E - (.*\n)* - \n\QGRANT SELECT ON TABLE public.t TO grant1;\E - \n\QGRANT INSERT ON TABLE public.t TO grant2;\E - \n\QGRANT ALL ON TABLE public.t TO grant3;\E - (.*\n)* - \n\QGRANT CREATE,CONNECT ON DATABASE tapgrantsdb TO grant4;\E - /xm, - }, - - excluding_databases => { - setup_sql => 'CREATE DATABASE db1; - \c db1 - CREATE TABLE t1 (id int); - INSERT INTO t1 VALUES (1), (2), (3), (4); - CREATE TABLE t2 (id int); - INSERT INTO t2 VALUES (1), (2), (3), (4); - - CREATE DATABASE db2; - \c db2 - CREATE TABLE t3 (id int); - INSERT INTO t3 VALUES (1), (2), (3), (4); - CREATE TABLE t4 (id int); - INSERT INTO t4 VALUES (1), (2), (3), (4); - - CREATE DATABASE dbex3; - \c dbex3 - CREATE TABLE t5 (id int); - INSERT INTO t5 VALUES (1), (2), (3), (4); - CREATE TABLE t6 (id int); - INSERT INTO t6 VALUES (1), (2), (3), (4); - - CREATE DATABASE dbex4; - \c dbex4 - CREATE TABLE t7 (id int); - INSERT INTO t7 VALUES (1), (2), (3), (4); - CREATE TABLE t8 (id int); - INSERT INTO t8 VALUES (1), (2), (3), (4); - - CREATE DATABASE db5; - \c db5 - CREATE TABLE t9 (id int); - INSERT INTO t9 VALUES (1), (2), (3), (4); - CREATE TABLE t10 (id int); - INSERT INTO t10 VALUES (1), (2), (3), (4); - ', - dump_cmd => [ - 'pg_dumpall', - '--format' => 'directory', - '--file' => "$tempdir/excluding_databases", - '--exclude-database' => 'dbex*', - ], - restore_cmd => [ - 'pg_restore', '-C', - '--format' => 'directory', - '--file' => "$tempdir/excluding_databases.sql", - '--exclude-database' => 'db5', - "$tempdir/excluding_databases", - ], - like => qr/^ - \n\QCREATE DATABASE db1\E - (.*\n)* - \n\QCREATE TABLE public.t1 (\E - (.*\n)* - \n\QCREATE TABLE public.t2 (\E - (.*\n)* - \n\QCREATE DATABASE db2\E - (.*\n)* - \n\QCREATE TABLE public.t3 (\E - (.*\n)* - \n\QCREATE TABLE public.t4 (/xm, - unlike => qr/^ - \n\QCREATE DATABASE db3\E - (.*\n)* - \n\QCREATE TABLE public.t5 (\E - (.*\n)* - \n\QCREATE TABLE public.t6 (\E - (.*\n)* - \n\QCREATE DATABASE db4\E - (.*\n)* - \n\QCREATE TABLE public.t7 (\E - (.*\n)* - \n\QCREATE TABLE public.t8 (\E - \n\QCREATE DATABASE db5\E - (.*\n)* - \n\QCREATE TABLE public.t9 (\E - (.*\n)* - \n\QCREATE TABLE public.t10 (\E - /xm, - }, - - format_directory => { - setup_sql => "CREATE TABLE format_directory(a int, b boolean, c text); - INSERT INTO format_directory VALUES (1, true, 'name1'), (2, false, 'name2');", - dump_cmd => [ - 'pg_dumpall', - '--format' => 'directory', - '--file' => "$tempdir/format_directory", - ], - restore_cmd => [ - 'pg_restore', '-C', - '--format' => 'directory', - '--file' => "$tempdir/format_directory.sql", - "$tempdir/format_directory", - ], - like => qr/^\n\QCOPY public.format_directory (a, b, c) FROM stdin;/xm - }, - - format_tar => { - setup_sql => "CREATE TABLE format_tar(a int, b boolean, c text); - INSERT INTO format_tar VALUES (1, false, 'name3'), (2, true, 'name4');", - dump_cmd => [ - 'pg_dumpall', - '--format' => 'tar', - '--file' => "$tempdir/format_tar", - ], - restore_cmd => [ - 'pg_restore', '-C', - '--format' => 'tar', - '--file' => "$tempdir/format_tar.sql", - "$tempdir/format_tar", - ], - like => qr/^\n\QCOPY public.format_tar (a, b, c) FROM stdin;/xm - }, - - format_custom => { - setup_sql => "CREATE TABLE format_custom(a int, b boolean, c text); - INSERT INTO format_custom VALUES (1, false, 'name5'), (2, true, 'name6');", - dump_cmd => [ - 'pg_dumpall', - '--format' => 'custom', - '--file' => "$tempdir/format_custom", - ], - restore_cmd => [ - 'pg_restore', '-C', - '--format' => 'custom', - '--file' => "$tempdir/format_custom.sql", - "$tempdir/format_custom", - ], - like => qr/^ \n\QCOPY public.format_custom (a, b, c) FROM stdin;/xm - }, - - dump_globals_only => { - setup_sql => "CREATE TABLE format_dir(a int, b boolean, c text); - INSERT INTO format_dir VALUES (1, false, 'name5'), (2, true, 'name6');", - dump_cmd => [ - 'pg_dumpall', - '--format' => 'directory', - '--globals-only', - '--file' => "$tempdir/dump_globals_only", - ], - restore_cmd => [ - 'pg_restore', '-C', '--globals-only', - '--format' => 'directory', - '--file' => "$tempdir/dump_globals_only.sql", - "$tempdir/dump_globals_only", - ], - like => qr/ - ^\s*\QCREATE ROLE dumpall;\E\s*\n - /xm - },); - -# First execute the setup_sql -foreach my $run (sort keys %pgdumpall_runs) -{ - if ($pgdumpall_runs{$run}->{setup_sql}) - { - $node->safe_psql($run_db, $pgdumpall_runs{$run}->{setup_sql}); - } -} - -# Execute the tests -foreach my $run (sort keys %pgdumpall_runs) -{ - # Create a new target cluster to pg_restore each test case run so that we - # don't need to take care of the cleanup from the target cluster after each - # run. - my $target_node = PostgreSQL::Test::Cluster->new("target_$run"); - $target_node->init; - $target_node->start; - - # Dumpall from node cluster. - $node->command_ok(\@{ $pgdumpall_runs{$run}->{dump_cmd} }, - "$run: pg_dumpall runs"); - - # Restore the dump on "target_node" cluster. - my @restore_cmd = ( - @{ $pgdumpall_runs{$run}->{restore_cmd} }, - '--host', $target_node->host, '--port', $target_node->port); - - my ($stdout, $stderr) = run_command(\@restore_cmd); - - # pg_restore --file output file. - my $output_file = slurp_file("$tempdir/${run}.sql"); - - if ( !($pgdumpall_runs{$run}->{like}) - && !($pgdumpall_runs{$run}->{unlike})) - { - die "missing \"like\" or \"unlike\" in test \"$run\""; - } - - if ($pgdumpall_runs{$run}->{like}) - { - like($output_file, $pgdumpall_runs{$run}->{like}, "should dump $run"); - } - - if ($pgdumpall_runs{$run}->{unlike}) - { - unlike( - $output_file, - $pgdumpall_runs{$run}->{unlike}, - "should not dump $run"); - } -} - -# Some negative test case with dump of pg_dumpall and restore using pg_restore -# test case 1: when -C is not used in pg_restore with dump of pg_dumpall -$node->command_fails_like( - [ - 'pg_restore', - "$tempdir/format_custom", - '--format' => 'custom', - '--file' => "$tempdir/error_test.sql", - ], - qr/\Qpg_restore: error: option -C\/--create must be specified when restoring an archive created by pg_dumpall\E/, - 'When -C is not used in pg_restore with dump of pg_dumpall'); - -# test case 2: When --list option is used with dump of pg_dumpall -$node->command_fails_like( - [ - 'pg_restore', - "$tempdir/format_custom", '-C', - '--format' => 'custom', - '--list', - '--file' => "$tempdir/error_test.sql", - ], - qr/\Qpg_restore: error: option -l\/--list cannot be used when restoring an archive created by pg_dumpall\E/, - 'When --list is used in pg_restore with dump of pg_dumpall'); - -# test case 3: When non-exist database is given with -d option -$node->command_fails_like( - [ - 'pg_restore', - "$tempdir/format_custom", '-C', - '--format' => 'custom', - '-d' => 'dbpq', - ], - qr/\Qpg_restore: error: could not connect to database "dbpq"\E/, - 'When non-existent database is given with -d option in pg_restore with dump of pg_dumpall' -); - -$node->stop('fast'); - -done_testing(); diff --git a/src/bin/pg_upgrade/check.c b/src/bin/pg_upgrade/check.c index 5e6403f07731b..310f53c55771b 100644 --- a/src/bin/pg_upgrade/check.c +++ b/src/bin/pg_upgrade/check.c @@ -956,12 +956,12 @@ check_for_new_tablespace_dir(void) prep_status("Checking for new cluster tablespace directories"); - for (tblnum = 0; tblnum < os_info.num_old_tablespaces; tblnum++) + for (tblnum = 0; tblnum < new_cluster.num_tablespaces; tblnum++) { struct stat statbuf; snprintf(new_tablespace_dir, MAXPGPATH, "%s%s", - os_info.old_tablespaces[tblnum], + new_cluster.tablespaces[tblnum], new_cluster.tablespace_suffix); if (stat(new_tablespace_dir, &statbuf) == 0 || errno != ENOENT) @@ -1013,17 +1013,17 @@ create_script_for_old_cluster_deletion(char **deletion_script_file_name) * directory. We can't create a proper old cluster delete script in that * case. */ - for (tblnum = 0; tblnum < os_info.num_old_tablespaces; tblnum++) + for (tblnum = 0; tblnum < new_cluster.num_tablespaces; tblnum++) { - char old_tablespace_dir[MAXPGPATH]; + char new_tablespace_dir[MAXPGPATH]; - strlcpy(old_tablespace_dir, os_info.old_tablespaces[tblnum], MAXPGPATH); - canonicalize_path(old_tablespace_dir); - if (path_is_prefix_of_path(old_cluster_pgdata, old_tablespace_dir)) + strlcpy(new_tablespace_dir, new_cluster.tablespaces[tblnum], MAXPGPATH); + canonicalize_path(new_tablespace_dir); + if (path_is_prefix_of_path(old_cluster_pgdata, new_tablespace_dir)) { /* reproduce warning from CREATE TABLESPACE that is in the log */ pg_log(PG_WARNING, - "\nWARNING: user-defined tablespace locations should not be inside the data directory, i.e. %s", old_tablespace_dir); + "\nWARNING: user-defined tablespace locations should not be inside the data directory, i.e. %s", new_tablespace_dir); /* Unlink file in case it is left over from a previous run. */ unlink(*deletion_script_file_name); @@ -1051,9 +1051,9 @@ create_script_for_old_cluster_deletion(char **deletion_script_file_name) /* delete old cluster's alternate tablespaces */ old_tblspc_suffix = pg_strdup(old_cluster.tablespace_suffix); fix_path_separator(old_tblspc_suffix); - for (tblnum = 0; tblnum < os_info.num_old_tablespaces; tblnum++) + for (tblnum = 0; tblnum < old_cluster.num_tablespaces; tblnum++) fprintf(script, RMDIR_CMD " %c%s%s%c\n", PATH_QUOTE, - fix_path_separator(os_info.old_tablespaces[tblnum]), + fix_path_separator(old_cluster.tablespaces[tblnum]), old_tblspc_suffix, PATH_QUOTE); pfree(old_tblspc_suffix); diff --git a/src/bin/pg_upgrade/dump.c b/src/bin/pg_upgrade/dump.c index 183f08ce1e86f..55f6e7b4d9c3e 100644 --- a/src/bin/pg_upgrade/dump.c +++ b/src/bin/pg_upgrade/dump.c @@ -58,7 +58,7 @@ generate_old_dump(void) (user_opts.transfer_mode == TRANSFER_MODE_SWAP) ? "" : "--sequence-data", log_opts.verbose ? "--verbose" : "", - user_opts.do_statistics ? "--with-statistics" : "--no-statistics", + user_opts.do_statistics ? "--statistics" : "--no-statistics", log_opts.dumpdir, sql_file_name, escaped_connstr.data); diff --git a/src/bin/pg_upgrade/info.c b/src/bin/pg_upgrade/info.c index a437067cdca82..c39eb077c2fae 100644 --- a/src/bin/pg_upgrade/info.c +++ b/src/bin/pg_upgrade/info.c @@ -443,10 +443,26 @@ get_db_infos(ClusterInfo *cluster) for (tupnum = 0; tupnum < ntups; tupnum++) { + char *spcloc = PQgetvalue(res, tupnum, i_spclocation); + bool inplace = spcloc[0] && !is_absolute_path(spcloc); + dbinfos[tupnum].db_oid = atooid(PQgetvalue(res, tupnum, i_oid)); dbinfos[tupnum].db_name = pg_strdup(PQgetvalue(res, tupnum, i_datname)); - snprintf(dbinfos[tupnum].db_tablespace, sizeof(dbinfos[tupnum].db_tablespace), "%s", - PQgetvalue(res, tupnum, i_spclocation)); + + /* + * The tablespace location might be "", meaning the cluster default + * location, i.e. pg_default or pg_global. For in-place tablespaces, + * pg_tablespace_location() returns a path relative to the data + * directory. + */ + if (inplace) + snprintf(dbinfos[tupnum].db_tablespace, + sizeof(dbinfos[tupnum].db_tablespace), + "%s/%s", cluster->pgdata, spcloc); + else + snprintf(dbinfos[tupnum].db_tablespace, + sizeof(dbinfos[tupnum].db_tablespace), + "%s", spcloc); } PQclear(res); @@ -616,11 +632,21 @@ process_rel_infos(DbInfo *dbinfo, PGresult *res, void *arg) /* Is the tablespace oid non-default? */ if (atooid(PQgetvalue(res, relnum, i_reltablespace)) != 0) { + char *spcloc = PQgetvalue(res, relnum, i_spclocation); + bool inplace = spcloc[0] && !is_absolute_path(spcloc); + /* * The tablespace location might be "", meaning the cluster - * default location, i.e. pg_default or pg_global. + * default location, i.e. pg_default or pg_global. For in-place + * tablespaces, pg_tablespace_location() returns a path relative + * to the data directory. */ - tablespace = PQgetvalue(res, relnum, i_spclocation); + if (inplace) + tablespace = psprintf("%s/%s", + os_info.running_cluster->pgdata, + spcloc); + else + tablespace = spcloc; /* Can we reuse the previous string allocation? */ if (last_tablespace && strcmp(tablespace, last_tablespace) == 0) @@ -630,6 +656,10 @@ process_rel_infos(DbInfo *dbinfo, PGresult *res, void *arg) last_tablespace = curr->tablespace = pg_strdup(tablespace); curr->tblsp_alloc = true; } + + /* Free palloc'd string for in-place tablespaces. */ + if (inplace) + pfree(tablespace); } else /* A zero reltablespace oid indicates the database tablespace. */ diff --git a/src/bin/pg_upgrade/parallel.c b/src/bin/pg_upgrade/parallel.c index 056aa2edaee3f..6d7941844a7c8 100644 --- a/src/bin/pg_upgrade/parallel.c +++ b/src/bin/pg_upgrade/parallel.c @@ -40,6 +40,7 @@ typedef struct char *old_pgdata; char *new_pgdata; char *old_tablespace; + char *new_tablespace; } transfer_thread_arg; static exec_thread_arg **exec_thread_args; @@ -171,7 +172,7 @@ win32_exec_prog(exec_thread_arg *args) void parallel_transfer_all_new_dbs(DbInfoArr *old_db_arr, DbInfoArr *new_db_arr, char *old_pgdata, char *new_pgdata, - char *old_tablespace) + char *old_tablespace, char *new_tablespace) { #ifndef WIN32 pid_t child; @@ -181,7 +182,7 @@ parallel_transfer_all_new_dbs(DbInfoArr *old_db_arr, DbInfoArr *new_db_arr, #endif if (user_opts.jobs <= 1) - transfer_all_new_dbs(old_db_arr, new_db_arr, old_pgdata, new_pgdata, NULL); + transfer_all_new_dbs(old_db_arr, new_db_arr, old_pgdata, new_pgdata, NULL, NULL); else { /* parallel */ @@ -225,7 +226,7 @@ parallel_transfer_all_new_dbs(DbInfoArr *old_db_arr, DbInfoArr *new_db_arr, if (child == 0) { transfer_all_new_dbs(old_db_arr, new_db_arr, old_pgdata, new_pgdata, - old_tablespace); + old_tablespace, new_tablespace); /* if we take another exit path, it will be non-zero */ /* use _exit to skip atexit() functions */ _exit(0); @@ -246,6 +247,7 @@ parallel_transfer_all_new_dbs(DbInfoArr *old_db_arr, DbInfoArr *new_db_arr, new_arg->new_pgdata = pg_strdup(new_pgdata); pg_free(new_arg->old_tablespace); new_arg->old_tablespace = old_tablespace ? pg_strdup(old_tablespace) : NULL; + new_arg->new_tablespace = new_tablespace ? pg_strdup(new_tablespace) : NULL; child = (HANDLE) _beginthreadex(NULL, 0, (void *) win32_transfer_all_new_dbs, new_arg, 0, NULL); @@ -263,7 +265,8 @@ DWORD win32_transfer_all_new_dbs(transfer_thread_arg *args) { transfer_all_new_dbs(args->old_db_arr, args->new_db_arr, args->old_pgdata, - args->new_pgdata, args->old_tablespace); + args->new_pgdata, args->old_tablespace, + args->new_tablespace); /* terminates thread */ return 0; diff --git a/src/bin/pg_upgrade/pg_upgrade.h b/src/bin/pg_upgrade/pg_upgrade.h index e9401430e697f..0ef47be0dc199 100644 --- a/src/bin/pg_upgrade/pg_upgrade.h +++ b/src/bin/pg_upgrade/pg_upgrade.h @@ -300,6 +300,8 @@ typedef struct uint32 major_version; /* PG_VERSION of cluster */ char major_version_str[64]; /* string PG_VERSION of cluster */ uint32 bin_version; /* version returned from pg_ctl */ + char **tablespaces; /* tablespace directories */ + int num_tablespaces; const char *tablespace_suffix; /* directory specification */ int nsubs; /* number of subscriptions */ bool sub_retain_dead_tuples; /* whether a subscription enables @@ -356,8 +358,6 @@ typedef struct const char *progname; /* complete pathname for this program */ char *user; /* username for clusters */ bool user_specified; /* user specified on command-line */ - char **old_tablespaces; /* tablespaces */ - int num_old_tablespaces; LibraryInfo *libraries; /* loadable libraries */ int num_libraries; ClusterInfo *running_cluster; @@ -457,7 +457,7 @@ void transfer_all_new_tablespaces(DbInfoArr *old_db_arr, DbInfoArr *new_db_arr, char *old_pgdata, char *new_pgdata); void transfer_all_new_dbs(DbInfoArr *old_db_arr, DbInfoArr *new_db_arr, char *old_pgdata, char *new_pgdata, - char *old_tablespace); + char *old_tablespace, char *new_tablespace); /* tablespace.c */ @@ -505,7 +505,7 @@ void parallel_exec_prog(const char *log_file, const char *opt_log_file, const char *fmt,...) pg_attribute_printf(3, 4); void parallel_transfer_all_new_dbs(DbInfoArr *old_db_arr, DbInfoArr *new_db_arr, char *old_pgdata, char *new_pgdata, - char *old_tablespace); + char *old_tablespace, char *new_tablespace); bool reap_child(bool wait_for_child); /* task.c */ diff --git a/src/bin/pg_upgrade/relfilenumber.c b/src/bin/pg_upgrade/relfilenumber.c index 8d8e816a01fa4..38c17ceabf222 100644 --- a/src/bin/pg_upgrade/relfilenumber.c +++ b/src/bin/pg_upgrade/relfilenumber.c @@ -17,7 +17,7 @@ #include "common/logging.h" #include "pg_upgrade.h" -static void transfer_single_new_db(FileNameMap *maps, int size, char *old_tablespace); +static void transfer_single_new_db(FileNameMap *maps, int size, char *old_tablespace, char *new_tablespace); static void transfer_relfile(FileNameMap *map, const char *type_suffix, bool vm_must_add_frozenbit); /* @@ -136,21 +136,22 @@ transfer_all_new_tablespaces(DbInfoArr *old_db_arr, DbInfoArr *new_db_arr, */ if (user_opts.jobs <= 1) parallel_transfer_all_new_dbs(old_db_arr, new_db_arr, old_pgdata, - new_pgdata, NULL); + new_pgdata, NULL, NULL); else { int tblnum; /* transfer default tablespace */ parallel_transfer_all_new_dbs(old_db_arr, new_db_arr, old_pgdata, - new_pgdata, old_pgdata); + new_pgdata, old_pgdata, new_pgdata); - for (tblnum = 0; tblnum < os_info.num_old_tablespaces; tblnum++) + for (tblnum = 0; tblnum < old_cluster.num_tablespaces; tblnum++) parallel_transfer_all_new_dbs(old_db_arr, new_db_arr, old_pgdata, new_pgdata, - os_info.old_tablespaces[tblnum]); + old_cluster.tablespaces[tblnum], + new_cluster.tablespaces[tblnum]); /* reap all children */ while (reap_child(true) == true) ; @@ -169,7 +170,8 @@ transfer_all_new_tablespaces(DbInfoArr *old_db_arr, DbInfoArr *new_db_arr, */ void transfer_all_new_dbs(DbInfoArr *old_db_arr, DbInfoArr *new_db_arr, - char *old_pgdata, char *new_pgdata, char *old_tablespace) + char *old_pgdata, char *new_pgdata, + char *old_tablespace, char *new_tablespace) { int old_dbnum, new_dbnum; @@ -204,7 +206,7 @@ transfer_all_new_dbs(DbInfoArr *old_db_arr, DbInfoArr *new_db_arr, new_pgdata); if (n_maps) { - transfer_single_new_db(mappings, n_maps, old_tablespace); + transfer_single_new_db(mappings, n_maps, old_tablespace, new_tablespace); } /* We allocate something even for n_maps == 0 */ pg_free(mappings); @@ -234,10 +236,10 @@ transfer_all_new_dbs(DbInfoArr *old_db_arr, DbInfoArr *new_db_arr, * moved_db_dir: Destination for the pg_restore-generated database directory. */ static bool -prepare_for_swap(const char *old_tablespace, Oid db_oid, - char *old_catalog_dir, char *new_db_dir, char *moved_db_dir) +prepare_for_swap(const char *old_tablespace, const char *new_tablespace, + Oid db_oid, char *old_catalog_dir, char *new_db_dir, + char *moved_db_dir) { - const char *new_tablespace; const char *old_tblspc_suffix; const char *new_tblspc_suffix; char old_tblspc[MAXPGPATH]; @@ -247,24 +249,14 @@ prepare_for_swap(const char *old_tablespace, Oid db_oid, struct stat st; if (strcmp(old_tablespace, old_cluster.pgdata) == 0) - { - new_tablespace = new_cluster.pgdata; - new_tblspc_suffix = "/base"; old_tblspc_suffix = "/base"; - } else - { - /* - * XXX: The below line is a hack to deal with the fact that we - * presently don't have an easy way to find the corresponding new - * tablespace's path. This will need to be fixed if/when we add - * pg_upgrade support for in-place tablespaces. - */ - new_tablespace = old_tablespace; + old_tblspc_suffix = old_cluster.tablespace_suffix; + if (strcmp(new_tablespace, new_cluster.pgdata) == 0) + new_tblspc_suffix = "/base"; + else new_tblspc_suffix = new_cluster.tablespace_suffix; - old_tblspc_suffix = old_cluster.tablespace_suffix; - } /* Old and new cluster paths. */ snprintf(old_tblspc, sizeof(old_tblspc), "%s%s", old_tablespace, old_tblspc_suffix); @@ -450,7 +442,7 @@ swap_catalog_files(FileNameMap *maps, int size, const char *old_catalog_dir, * during pg_restore. */ static void -do_swap(FileNameMap *maps, int size, char *old_tablespace) +do_swap(FileNameMap *maps, int size, char *old_tablespace, char *new_tablespace) { char old_catalog_dir[MAXPGPATH]; char new_db_dir[MAXPGPATH]; @@ -470,21 +462,23 @@ do_swap(FileNameMap *maps, int size, char *old_tablespace) */ if (old_tablespace) { - if (prepare_for_swap(old_tablespace, maps[0].db_oid, + if (prepare_for_swap(old_tablespace, new_tablespace, maps[0].db_oid, old_catalog_dir, new_db_dir, moved_db_dir)) swap_catalog_files(maps, size, old_catalog_dir, new_db_dir, moved_db_dir); } else { - if (prepare_for_swap(old_cluster.pgdata, maps[0].db_oid, + if (prepare_for_swap(old_cluster.pgdata, new_cluster.pgdata, maps[0].db_oid, old_catalog_dir, new_db_dir, moved_db_dir)) swap_catalog_files(maps, size, old_catalog_dir, new_db_dir, moved_db_dir); - for (int tblnum = 0; tblnum < os_info.num_old_tablespaces; tblnum++) + for (int tblnum = 0; tblnum < old_cluster.num_tablespaces; tblnum++) { - if (prepare_for_swap(os_info.old_tablespaces[tblnum], maps[0].db_oid, + if (prepare_for_swap(old_cluster.tablespaces[tblnum], + new_cluster.tablespaces[tblnum], + maps[0].db_oid, old_catalog_dir, new_db_dir, moved_db_dir)) swap_catalog_files(maps, size, old_catalog_dir, new_db_dir, moved_db_dir); @@ -498,7 +492,8 @@ do_swap(FileNameMap *maps, int size, char *old_tablespace) * create links for mappings stored in "maps" array. */ static void -transfer_single_new_db(FileNameMap *maps, int size, char *old_tablespace) +transfer_single_new_db(FileNameMap *maps, int size, + char *old_tablespace, char *new_tablespace) { int mapnum; bool vm_must_add_frozenbit = false; @@ -520,7 +515,7 @@ transfer_single_new_db(FileNameMap *maps, int size, char *old_tablespace) */ Assert(!vm_must_add_frozenbit); - do_swap(maps, size, old_tablespace); + do_swap(maps, size, old_tablespace, new_tablespace); return; } diff --git a/src/bin/pg_upgrade/t/006_transfer_modes.pl b/src/bin/pg_upgrade/t/006_transfer_modes.pl index 58fe8a8c7dcea..348f402146234 100644 --- a/src/bin/pg_upgrade/t/006_transfer_modes.pl +++ b/src/bin/pg_upgrade/t/006_transfer_modes.pl @@ -38,6 +38,13 @@ sub test_mode } $new->init(); + # allow_in_place_tablespaces is available as far back as v10. + if ($old->pg_version >= 10) + { + $new->append_conf('postgresql.conf', "allow_in_place_tablespaces = true"); + $old->append_conf('postgresql.conf', "allow_in_place_tablespaces = true"); + } + # Create a small variety of simple test objects on the old cluster. We'll # check that these reach the new version after upgrading. $old->start; @@ -49,8 +56,7 @@ sub test_mode $old->safe_psql('testdb1', "VACUUM FULL test2"); $old->safe_psql('testdb1', "CREATE SEQUENCE testseq START 5432"); - # For cross-version tests, we can also check that pg_upgrade handles - # tablespaces. + # If an old installation is provided, we can test non-in-place tablespaces. if (defined($ENV{oldinstall})) { my $tblspc = PostgreSQL::Test::Utils::tempdir_short(); @@ -64,6 +70,19 @@ sub test_mode $old->safe_psql('testdb2', "CREATE TABLE test4 AS SELECT generate_series(400, 502)"); } + + # If the old cluster is >= v10, we can test in-place tablespaces. + if ($old->pg_version >= 10) + { + $old->safe_psql('postgres', + "CREATE TABLESPACE inplc_tblspc LOCATION ''"); + $old->safe_psql('postgres', + "CREATE DATABASE testdb3 TABLESPACE inplc_tblspc"); + $old->safe_psql('postgres', + "CREATE TABLE test5 TABLESPACE inplc_tblspc AS SELECT generate_series(503, 606)"); + $old->safe_psql('testdb3', + "CREATE TABLE test6 AS SELECT generate_series(607, 711)"); + } $old->stop; my $result = command_ok_or_fails_like( @@ -94,8 +113,7 @@ sub test_mode $result = $new->safe_psql('testdb1', "SELECT nextval('testseq')"); is($result, '5432', "sequence data after pg_upgrade $mode"); - # For cross-version tests, we should have some objects in a non-default - # tablespace. + # Tests for non-in-place tablespaces. if (defined($ENV{oldinstall})) { $result = @@ -105,6 +123,15 @@ sub test_mode $new->safe_psql('testdb2', "SELECT COUNT(*) FROM test4"); is($result, '103', "test4 data after pg_upgrade $mode"); } + + # Tests for in-place tablespaces. + if ($old->pg_version >= 10) + { + $result = $new->safe_psql('postgres', "SELECT COUNT(*) FROM test5"); + is($result, '104', "test5 data after pg_upgrade $mode"); + $result = $new->safe_psql('testdb3', "SELECT COUNT(*) FROM test6"); + is($result, '105', "test6 data after pg_upgrade $mode"); + } $new->stop; } diff --git a/src/bin/pg_upgrade/tablespace.c b/src/bin/pg_upgrade/tablespace.c index 3520a75ba317d..151d74e17349b 100644 --- a/src/bin/pg_upgrade/tablespace.c +++ b/src/bin/pg_upgrade/tablespace.c @@ -23,10 +23,20 @@ init_tablespaces(void) set_tablespace_directory_suffix(&old_cluster); set_tablespace_directory_suffix(&new_cluster); - if (os_info.num_old_tablespaces > 0 && + if (old_cluster.num_tablespaces > 0 && strcmp(old_cluster.tablespace_suffix, new_cluster.tablespace_suffix) == 0) - pg_fatal("Cannot upgrade to/from the same system catalog version when\n" - "using tablespaces."); + { + for (int i = 0; i < old_cluster.num_tablespaces; i++) + { + /* + * In-place tablespaces are okay for same-version upgrades because + * their paths will differ between clusters. + */ + if (strcmp(old_cluster.tablespaces[i], new_cluster.tablespaces[i]) == 0) + pg_fatal("Cannot upgrade to/from the same system catalog version when\n" + "using tablespaces."); + } + } } @@ -53,19 +63,48 @@ get_tablespace_paths(void) res = executeQueryOrDie(conn, "%s", query); - if ((os_info.num_old_tablespaces = PQntuples(res)) != 0) - os_info.old_tablespaces = - (char **) pg_malloc(os_info.num_old_tablespaces * sizeof(char *)); + old_cluster.num_tablespaces = PQntuples(res); + new_cluster.num_tablespaces = PQntuples(res); + + if (PQntuples(res) != 0) + { + old_cluster.tablespaces = + (char **) pg_malloc(old_cluster.num_tablespaces * sizeof(char *)); + new_cluster.tablespaces = + (char **) pg_malloc(new_cluster.num_tablespaces * sizeof(char *)); + } else - os_info.old_tablespaces = NULL; + { + old_cluster.tablespaces = NULL; + new_cluster.tablespaces = NULL; + } i_spclocation = PQfnumber(res, "spclocation"); - for (tblnum = 0; tblnum < os_info.num_old_tablespaces; tblnum++) + for (tblnum = 0; tblnum < old_cluster.num_tablespaces; tblnum++) { struct stat statBuf; + char *spcloc = PQgetvalue(res, tblnum, i_spclocation); - os_info.old_tablespaces[tblnum] = pg_strdup(PQgetvalue(res, tblnum, i_spclocation)); + /* + * For now, we do not expect non-in-place tablespaces to move during + * upgrade. If that changes, it will likely become necessary to run + * the above query on the new cluster, too. + * + * pg_tablespace_location() returns absolute paths for non-in-place + * tablespaces and relative paths for in-place ones, so we use + * is_absolute_path() to distinguish between them. + */ + if (is_absolute_path(PQgetvalue(res, tblnum, i_spclocation))) + { + old_cluster.tablespaces[tblnum] = pg_strdup(spcloc); + new_cluster.tablespaces[tblnum] = old_cluster.tablespaces[tblnum]; + } + else + { + old_cluster.tablespaces[tblnum] = psprintf("%s/%s", old_cluster.pgdata, spcloc); + new_cluster.tablespaces[tblnum] = psprintf("%s/%s", new_cluster.pgdata, spcloc); + } /* * Check that the tablespace path exists and is a directory. @@ -76,21 +115,21 @@ get_tablespace_paths(void) * that contains user tablespaces is moved as part of pg_upgrade * preparation and the symbolic links are not updated. */ - if (stat(os_info.old_tablespaces[tblnum], &statBuf) != 0) + if (stat(old_cluster.tablespaces[tblnum], &statBuf) != 0) { if (errno == ENOENT) report_status(PG_FATAL, "tablespace directory \"%s\" does not exist", - os_info.old_tablespaces[tblnum]); + old_cluster.tablespaces[tblnum]); else report_status(PG_FATAL, "could not stat tablespace directory \"%s\": %m", - os_info.old_tablespaces[tblnum]); + old_cluster.tablespaces[tblnum]); } if (!S_ISDIR(statBuf.st_mode)) report_status(PG_FATAL, "tablespace path \"%s\" is not a directory", - os_info.old_tablespaces[tblnum]); + old_cluster.tablespaces[tblnum]); } PQclear(res); diff --git a/src/bin/pgbench/pgbench.c b/src/bin/pgbench/pgbench.c index 497a936c141f3..125f3c7bbbe5b 100644 --- a/src/bin/pgbench/pgbench.c +++ b/src/bin/pgbench/pgbench.c @@ -3495,6 +3495,8 @@ doRetry(CState *st, pg_time_usec_t *now) static int discardUntilSync(CState *st) { + bool received_sync = false; + /* send a sync */ if (!PQpipelineSync(st->con)) { @@ -3509,10 +3511,21 @@ discardUntilSync(CState *st) PGresult *res = PQgetResult(st->con); if (PQresultStatus(res) == PGRES_PIPELINE_SYNC) + received_sync = true; + else if (received_sync) { - PQclear(res); - res = PQgetResult(st->con); + /* + * PGRES_PIPELINE_SYNC must be followed by another + * PGRES_PIPELINE_SYNC or NULL; otherwise, assert failure. + */ Assert(res == NULL); + + /* + * Reset ongoing sync count to 0 since all PGRES_PIPELINE_SYNC + * results have been discarded. + */ + st->num_syncs = 0; + PQclear(res); break; } PQclear(res); diff --git a/src/bin/psql/tab-complete.in.c b/src/bin/psql/tab-complete.in.c index dbc586c5bc370..1f2ca946fc500 100644 --- a/src/bin/psql/tab-complete.in.c +++ b/src/bin/psql/tab-complete.in.c @@ -1010,7 +1010,7 @@ static const SchemaQuery Query_for_trigger_of_table = { #define Query_for_list_of_database_vars \ "SELECT conf FROM ("\ -" SELECT setdatabase, pg_catalog.split_part(unnest(setconfig),'=',1) conf"\ +" SELECT setdatabase, pg_catalog.split_part(pg_catalog.unnest(setconfig),'=',1) conf"\ " FROM pg_db_role_setting "\ " ) s, pg_database d "\ " WHERE s.setdatabase = d.oid "\ @@ -1086,9 +1086,12 @@ Keywords_for_list_of_owner_roles, "PUBLIC" " WHERE usename LIKE '%s'" #define Query_for_list_of_user_vars \ -" SELECT pg_catalog.split_part(pg_catalog.unnest(rolconfig),'=',1) "\ -" FROM pg_catalog.pg_roles "\ -" WHERE rolname LIKE '%s'" +"SELECT conf FROM ("\ +" SELECT rolname, pg_catalog.split_part(pg_catalog.unnest(rolconfig),'=',1) conf"\ +" FROM pg_catalog.pg_roles"\ +" ) s"\ +" WHERE s.conf like '%s' "\ +" AND s.rolname LIKE '%s'" #define Query_for_list_of_access_methods \ " SELECT amname "\ @@ -2517,7 +2520,10 @@ match_previous_words(int pattern_id, /* ALTER USER,ROLE RESET */ else if (Matches("ALTER", "USER|ROLE", MatchAny, "RESET")) + { + set_completion_reference(prev2_wd); COMPLETE_WITH_QUERY_PLUS(Query_for_list_of_user_vars, "ALL"); + } /* ALTER USER,ROLE WITH */ else if (Matches("ALTER", "USER|ROLE", MatchAny, "WITH")) @@ -5015,7 +5021,7 @@ match_previous_words(int pattern_id, /* Complete with a variable name */ else if (TailMatches("SET|RESET") && !TailMatches("UPDATE", MatchAny, "SET") && - !TailMatches("ALTER", "DATABASE", MatchAny, "RESET")) + !TailMatches("ALTER", "DATABASE|USER|ROLE", MatchAny, "RESET")) COMPLETE_WITH_QUERY_VERBATIM_PLUS(Query_for_list_of_set_vars, "CONSTRAINTS", "TRANSACTION", diff --git a/src/common/Makefile b/src/common/Makefile index 1e2b91c83c4c4..2c720caa50972 100644 --- a/src/common/Makefile +++ b/src/common/Makefile @@ -163,7 +163,7 @@ libpgcommon_shlib.a: $(OBJS_SHLIB) # The JSON API normally exits on out-of-memory; disable that behavior for shared # library builds. This requires libpq's pqexpbuffer.h. jsonapi_shlib.o: override CPPFLAGS += -DJSONAPI_USE_PQEXPBUFFER -jsonapi_shlib.o: override CPPFLAGS += -I$(libpq_srcdir) +jsonapi_shlib.o: override CPPFLAGS := -I$(libpq_srcdir) $(CPPFLAGS) # Because this uses its own compilation rule, it doesn't use the # dependency tracking logic from Makefile.global. To make sure that diff --git a/src/include/catalog/pg_subscription_rel.h b/src/include/catalog/pg_subscription_rel.h index c91797c869c24..f458447a0e5fb 100644 --- a/src/include/catalog/pg_subscription_rel.h +++ b/src/include/catalog/pg_subscription_rel.h @@ -85,7 +85,7 @@ typedef struct SubscriptionRelState extern void AddSubscriptionRelState(Oid subid, Oid relid, char state, XLogRecPtr sublsn, bool retain_lock); extern void UpdateSubscriptionRelState(Oid subid, Oid relid, char state, - XLogRecPtr sublsn); + XLogRecPtr sublsn, bool already_locked); extern char GetSubscriptionRelState(Oid subid, Oid relid, XLogRecPtr *sublsn); extern void RemoveSubscriptionRel(Oid subid, Oid relid); diff --git a/src/include/libpq/libpq-be-fe-helpers.h b/src/include/libpq/libpq-be-fe-helpers.h index af13bd6bf3da3..1c4a342090c33 100644 --- a/src/include/libpq/libpq-be-fe-helpers.h +++ b/src/include/libpq/libpq-be-fe-helpers.h @@ -30,17 +30,7 @@ #ifndef LIBPQ_BE_FE_HELPERS_H #define LIBPQ_BE_FE_HELPERS_H -/* - * Despite the name, BUILDING_DLL is set only when building code directly part - * of the backend. Which also is where libpq isn't allowed to be - * used. Obviously this doesn't protect against libpq-fe.h getting included - * otherwise, but perhaps still protects against a few mistakes... - */ -#ifdef BUILDING_DLL -#error "libpq may not be used code directly built into the backend" -#endif - -#include "libpq-fe.h" +#include "libpq/libpq-be-fe.h" #include "miscadmin.h" #include "storage/fd.h" #include "storage/latch.h" @@ -289,41 +279,30 @@ libpqsrv_exec_params(PGconn *conn, static inline PGresult * libpqsrv_get_result_last(PGconn *conn, uint32 wait_event_info) { - PGresult *volatile lastResult = NULL; + PGresult *lastResult = NULL; - /* In what follows, do not leak any PGresults on an error. */ - PG_TRY(); + for (;;) { - for (;;) - { - /* Wait for, and collect, the next PGresult. */ - PGresult *result; + /* Wait for, and collect, the next PGresult. */ + PGresult *result; - result = libpqsrv_get_result(conn, wait_event_info); - if (result == NULL) - break; /* query is complete, or failure */ + result = libpqsrv_get_result(conn, wait_event_info); + if (result == NULL) + break; /* query is complete, or failure */ - /* - * Emulate PQexec()'s behavior of returning the last result when - * there are many. - */ - PQclear(lastResult); - lastResult = result; - - if (PQresultStatus(lastResult) == PGRES_COPY_IN || - PQresultStatus(lastResult) == PGRES_COPY_OUT || - PQresultStatus(lastResult) == PGRES_COPY_BOTH || - PQstatus(conn) == CONNECTION_BAD) - break; - } - } - PG_CATCH(); - { + /* + * Emulate PQexec()'s behavior of returning the last result when there + * are many. + */ PQclear(lastResult); - PG_RE_THROW(); - } - PG_END_TRY(); + lastResult = result; + if (PQresultStatus(lastResult) == PGRES_COPY_IN || + PQresultStatus(lastResult) == PGRES_COPY_OUT || + PQresultStatus(lastResult) == PGRES_COPY_BOTH || + PQstatus(conn) == CONNECTION_BAD) + break; + } return lastResult; } @@ -462,13 +441,21 @@ exit: ; * This function is intended to be set via PQsetNoticeReceiver() so that * NOTICE, WARNING, and similar messages from the connection are reported via * ereport(), instead of being printed to stderr. + * + * Because this will be called from libpq with a "real" (not wrapped) + * PGresult, we need to temporarily ignore libpq-be-fe.h's wrapper macros + * for PGresult and also PQresultErrorMessage, and put back the wrappers + * afterwards. That's not pretty, but there seems no better alternative. */ +#undef PGresult +#undef PQresultErrorMessage + static inline void libpqsrv_notice_receiver(void *arg, const PGresult *res) { - char *message; + const char *message; int len; - char *prefix = (char *) arg; + const char *prefix = (const char *) arg; /* * Trim the trailing newline from the message text returned from @@ -484,4 +471,7 @@ libpqsrv_notice_receiver(void *arg, const PGresult *res) errmsg_internal("%s: %.*s", prefix, len, message)); } +#define PGresult libpqsrv_PGresult +#define PQresultErrorMessage libpqsrv_PQresultErrorMessage + #endif /* LIBPQ_BE_FE_HELPERS_H */ diff --git a/src/include/libpq/libpq-be-fe.h b/src/include/libpq/libpq-be-fe.h new file mode 100644 index 0000000000000..e3f796b023092 --- /dev/null +++ b/src/include/libpq/libpq-be-fe.h @@ -0,0 +1,259 @@ +/*------------------------------------------------------------------------- + * + * libpq-be-fe.h + * Wrapper functions for using libpq in extensions + * + * Code built directly into the backend is not allowed to link to libpq + * directly. Extension code is allowed to use libpq however. One of the + * main risks in doing so is leaking the malloc-allocated structures + * returned by libpq, causing a process-lifespan memory leak. + * + * This file provides wrapper objects to help in building memory-safe code. + * A PGresult object wrapped this way acts much as if it were palloc'd: + * it will go away when the specified context is reset or deleted. + * We might later extend the concept to other objects such as PGconns. + * + * See also the libpq-be-fe-helpers.h file, which provides additional + * facilities built on top of this one. + * + * Portions Copyright (c) 1996-2025, PostgreSQL Global Development Group + * Portions Copyright (c) 1994, Regents of the University of California + * + * src/include/libpq/libpq-be-fe.h + * + *------------------------------------------------------------------------- + */ +#ifndef LIBPQ_BE_FE_H +#define LIBPQ_BE_FE_H + +/* + * Despite the name, BUILDING_DLL is set only when building code directly part + * of the backend. Which also is where libpq isn't allowed to be + * used. Obviously this doesn't protect against libpq-fe.h getting included + * otherwise, but perhaps still protects against a few mistakes... + */ +#ifdef BUILDING_DLL +#error "libpq may not be used in code directly built into the backend" +#endif + +#include "libpq-fe.h" + +/* + * Memory-context-safe wrapper object for a PGresult. + */ +typedef struct libpqsrv_PGresult +{ + PGresult *res; /* the wrapped PGresult */ + MemoryContext ctx; /* the MemoryContext it's attached to */ + MemoryContextCallback cb; /* the callback that implements freeing */ +} libpqsrv_PGresult; + + +/* + * Wrap the given PGresult in a libpqsrv_PGresult object, so that it will + * go away automatically if the current memory context is reset or deleted. + * + * To avoid potential memory leaks, backend code must always apply this + * immediately to the output of any PGresult-yielding libpq function. + */ +static inline libpqsrv_PGresult * +libpqsrv_PQwrap(PGresult *res) +{ + libpqsrv_PGresult *bres; + MemoryContext ctx = CurrentMemoryContext; + + /* We pass through a NULL result as-is, since there's nothing to free */ + if (res == NULL) + return NULL; + /* Attempt to allocate the wrapper ... this had better not throw error */ + bres = (libpqsrv_PGresult *) + MemoryContextAllocExtended(ctx, + sizeof(libpqsrv_PGresult), + MCXT_ALLOC_NO_OOM); + /* If we failed to allocate a wrapper, free the PGresult before failing */ + if (bres == NULL) + { + PQclear(res); + ereport(ERROR, + (errcode(ERRCODE_OUT_OF_MEMORY), + errmsg("out of memory"))); + } + /* Okay, set up the wrapper */ + bres->res = res; + bres->ctx = ctx; + bres->cb.func = (MemoryContextCallbackFunction) PQclear; + bres->cb.arg = res; + MemoryContextRegisterResetCallback(ctx, &bres->cb); + return bres; +} + +/* + * Free a wrapped PGresult, after detaching it from the memory context. + * Like PQclear(), allow the argument to be NULL. + */ +static inline void +libpqsrv_PQclear(libpqsrv_PGresult *bres) +{ + if (bres) + { + MemoryContextUnregisterResetCallback(bres->ctx, &bres->cb); + PQclear(bres->res); + pfree(bres); + } +} + +/* + * Move a wrapped PGresult to have a different parent context. + */ +static inline libpqsrv_PGresult * +libpqsrv_PGresultSetParent(libpqsrv_PGresult *bres, MemoryContext ctx) +{ + libpqsrv_PGresult *newres; + + /* We pass through a NULL result as-is */ + if (bres == NULL) + return NULL; + /* Make a new wrapper in the target context, raising error on OOM */ + newres = (libpqsrv_PGresult *) + MemoryContextAlloc(ctx, sizeof(libpqsrv_PGresult)); + /* Okay, set up the new wrapper */ + newres->res = bres->res; + newres->ctx = ctx; + newres->cb.func = (MemoryContextCallbackFunction) PQclear; + newres->cb.arg = bres->res; + MemoryContextRegisterResetCallback(ctx, &newres->cb); + /* Disarm and delete the old wrapper */ + MemoryContextUnregisterResetCallback(bres->ctx, &bres->cb); + pfree(bres); + return newres; +} + +/* + * Convenience wrapper for PQgetResult. + * + * We could supply wrappers for other PGresult-returning functions too, + * but at present there's no need. + */ +static inline libpqsrv_PGresult * +libpqsrv_PQgetResult(PGconn *conn) +{ + return libpqsrv_PQwrap(PQgetResult(conn)); +} + +/* + * Accessor functions for libpqsrv_PGresult. While it's not necessary to use + * these, they emulate the behavior of the underlying libpq functions when + * passed a NULL pointer. This is particularly important for PQresultStatus, + * which is often the first check on a result. + */ + +static inline ExecStatusType +libpqsrv_PQresultStatus(const libpqsrv_PGresult *res) +{ + if (!res) + return PGRES_FATAL_ERROR; + return PQresultStatus(res->res); +} + +static inline const char * +libpqsrv_PQresultErrorMessage(const libpqsrv_PGresult *res) +{ + if (!res) + return ""; + return PQresultErrorMessage(res->res); +} + +static inline char * +libpqsrv_PQresultErrorField(const libpqsrv_PGresult *res, int fieldcode) +{ + if (!res) + return NULL; + return PQresultErrorField(res->res, fieldcode); +} + +static inline char * +libpqsrv_PQcmdStatus(const libpqsrv_PGresult *res) +{ + if (!res) + return NULL; + return PQcmdStatus(res->res); +} + +static inline int +libpqsrv_PQntuples(const libpqsrv_PGresult *res) +{ + if (!res) + return 0; + return PQntuples(res->res); +} + +static inline int +libpqsrv_PQnfields(const libpqsrv_PGresult *res) +{ + if (!res) + return 0; + return PQnfields(res->res); +} + +static inline char * +libpqsrv_PQgetvalue(const libpqsrv_PGresult *res, int tup_num, int field_num) +{ + if (!res) + return NULL; + return PQgetvalue(res->res, tup_num, field_num); +} + +static inline int +libpqsrv_PQgetlength(const libpqsrv_PGresult *res, int tup_num, int field_num) +{ + if (!res) + return 0; + return PQgetlength(res->res, tup_num, field_num); +} + +static inline int +libpqsrv_PQgetisnull(const libpqsrv_PGresult *res, int tup_num, int field_num) +{ + if (!res) + return 1; /* pretend it is null */ + return PQgetisnull(res->res, tup_num, field_num); +} + +static inline char * +libpqsrv_PQfname(const libpqsrv_PGresult *res, int field_num) +{ + if (!res) + return NULL; + return PQfname(res->res, field_num); +} + +static inline const char * +libpqsrv_PQcmdTuples(const libpqsrv_PGresult *res) +{ + if (!res) + return ""; + return PQcmdTuples(res->res); +} + +/* + * Redefine these libpq entry point names concerned with PGresults so that + * they will operate on libpqsrv_PGresults instead. This avoids needing to + * convert a lot of pre-existing code, and reduces the notational differences + * between frontend and backend libpq-using code. + */ +#define PGresult libpqsrv_PGresult +#define PQclear libpqsrv_PQclear +#define PQgetResult libpqsrv_PQgetResult +#define PQresultStatus libpqsrv_PQresultStatus +#define PQresultErrorMessage libpqsrv_PQresultErrorMessage +#define PQresultErrorField libpqsrv_PQresultErrorField +#define PQcmdStatus libpqsrv_PQcmdStatus +#define PQntuples libpqsrv_PQntuples +#define PQnfields libpqsrv_PQnfields +#define PQgetvalue libpqsrv_PQgetvalue +#define PQgetlength libpqsrv_PQgetlength +#define PQgetisnull libpqsrv_PQgetisnull +#define PQfname libpqsrv_PQfname +#define PQcmdTuples libpqsrv_PQcmdTuples + +#endif /* LIBPQ_BE_FE_H */ diff --git a/src/include/nodes/pathnodes.h b/src/include/nodes/pathnodes.h index e5dd15098f635..ad2726f026f7d 100644 --- a/src/include/nodes/pathnodes.h +++ b/src/include/nodes/pathnodes.h @@ -2133,10 +2133,12 @@ typedef struct MemoizePath * complete after caching the first record. */ bool binary_mode; /* true when cache key should be compared bit * by bit, false when using hash equality ops */ - Cardinality calls; /* expected number of rescans */ uint32 est_entries; /* The maximum number of entries that the * planner expects will fit in the cache, or 0 * if unknown */ + Cardinality est_calls; /* expected number of rescans */ + Cardinality est_unique_keys; /* estimated unique keys, for EXPLAIN */ + double est_hit_ratio; /* estimated cache hit ratio, for EXPLAIN */ } MemoizePath; /* diff --git a/src/include/nodes/plannodes.h b/src/include/nodes/plannodes.h index 4f59e30d62d5e..29d7732d6a031 100644 --- a/src/include/nodes/plannodes.h +++ b/src/include/nodes/plannodes.h @@ -28,6 +28,21 @@ * ---------------------------------------------------------------- */ +/* ---------------- + * PlannedStmtOrigin + * + * PlannedStmtOrigin identifies from where a PlannedStmt comes from. + * ---------------- + */ +typedef enum PlannedStmtOrigin +{ + PLAN_STMT_UNKNOWN = 0, /* plan origin is not yet known */ + PLAN_STMT_INTERNAL, /* generated internally by a query */ + PLAN_STMT_STANDARD, /* standard planned statement */ + PLAN_STMT_CACHE_GENERIC, /* Generic cached plan */ + PLAN_STMT_CACHE_CUSTOM, /* Custom cached plan */ +} PlannedStmtOrigin; + /* ---------------- * PlannedStmt node * @@ -58,6 +73,9 @@ typedef struct PlannedStmt /* plan identifier (can be set by plugins) */ int64 planId; + /* origin of plan */ + PlannedStmtOrigin planOrigin; + /* is it insert|update|delete|merge RETURNING? */ bool hasReturning; @@ -1056,6 +1074,16 @@ typedef struct Memoize /* paramids from param_exprs */ Bitmapset *keyparamids; + + /* Estimated number of rescans, for EXPLAIN */ + Cardinality est_calls; + + /* Estimated number of distinct lookup keys, for EXPLAIN */ + Cardinality est_unique_keys; + + /* Estimated cache hit ratio, for EXPLAIN */ + double est_hit_ratio; + } Memoize; /* ---------------- diff --git a/src/include/optimizer/pathnode.h b/src/include/optimizer/pathnode.h index 60dcdb77e41be..58936e963cb6b 100644 --- a/src/include/optimizer/pathnode.h +++ b/src/include/optimizer/pathnode.h @@ -90,7 +90,7 @@ extern MemoizePath *create_memoize_path(PlannerInfo *root, List *hash_operators, bool singlerow, bool binary_mode, - double calls); + Cardinality est_calls); extern UniquePath *create_unique_path(PlannerInfo *root, RelOptInfo *rel, Path *subpath, SpecialJoinInfo *sjinfo); extern GatherPath *create_gather_path(PlannerInfo *root, diff --git a/src/include/port/solaris.h b/src/include/port/solaris.h index e63a3bd824d6d..8ff40007c7f6a 100644 --- a/src/include/port/solaris.h +++ b/src/include/port/solaris.h @@ -24,3 +24,12 @@ #if defined(__i386__) #include #endif + +/* + * On original Solaris, PAM conversation procs lack a "const" in their + * declaration; but recent OpenIndiana versions put it there by default. + * The least messy way to deal with this is to define _PAM_LEGACY_NONCONST, + * which causes OpenIndiana to declare pam_conv per the Solaris tradition, + * and also use that symbol to control omitting the "const" in our own code. + */ +#define _PAM_LEGACY_NONCONST 1 diff --git a/src/include/storage/lwlock.h b/src/include/storage/lwlock.h index 08a72569ae5fd..5e717765764f4 100644 --- a/src/include/storage/lwlock.h +++ b/src/include/storage/lwlock.h @@ -176,51 +176,23 @@ extern void LWLockInitialize(LWLock *lock, int tranche_id); * Every tranche ID less than NUM_INDIVIDUAL_LWLOCKS is reserved; also, * we reserve additional tranche IDs for builtin tranches not included in * the set of individual LWLocks. A call to LWLockNewTrancheId will never - * return a value less than LWTRANCHE_FIRST_USER_DEFINED. + * return a value less than LWTRANCHE_FIRST_USER_DEFINED. The actual list of + * built-in tranches is kept in lwlocklist.h. */ typedef enum BuiltinTrancheIds { - LWTRANCHE_XACT_BUFFER = NUM_INDIVIDUAL_LWLOCKS, - LWTRANCHE_COMMITTS_BUFFER, - LWTRANCHE_SUBTRANS_BUFFER, - LWTRANCHE_MULTIXACTOFFSET_BUFFER, - LWTRANCHE_MULTIXACTMEMBER_BUFFER, - LWTRANCHE_NOTIFY_BUFFER, - LWTRANCHE_SERIAL_BUFFER, - LWTRANCHE_WAL_INSERT, - LWTRANCHE_BUFFER_CONTENT, - LWTRANCHE_REPLICATION_ORIGIN_STATE, - LWTRANCHE_REPLICATION_SLOT_IO, - LWTRANCHE_LOCK_FASTPATH, - LWTRANCHE_BUFFER_MAPPING, - LWTRANCHE_LOCK_MANAGER, - LWTRANCHE_PREDICATE_LOCK_MANAGER, - LWTRANCHE_PARALLEL_HASH_JOIN, - LWTRANCHE_PARALLEL_BTREE_SCAN, - LWTRANCHE_PARALLEL_QUERY_DSA, - LWTRANCHE_PER_SESSION_DSA, - LWTRANCHE_PER_SESSION_RECORD_TYPE, - LWTRANCHE_PER_SESSION_RECORD_TYPMOD, - LWTRANCHE_SHARED_TUPLESTORE, - LWTRANCHE_SHARED_TIDBITMAP, - LWTRANCHE_PARALLEL_APPEND, - LWTRANCHE_PER_XACT_PREDICATE_LIST, - LWTRANCHE_PGSTATS_DSA, - LWTRANCHE_PGSTATS_HASH, - LWTRANCHE_PGSTATS_DATA, - LWTRANCHE_LAUNCHER_DSA, - LWTRANCHE_LAUNCHER_HASH, - LWTRANCHE_DSM_REGISTRY_DSA, - LWTRANCHE_DSM_REGISTRY_HASH, - LWTRANCHE_COMMITTS_SLRU, - LWTRANCHE_MULTIXACTMEMBER_SLRU, - LWTRANCHE_MULTIXACTOFFSET_SLRU, - LWTRANCHE_NOTIFY_SLRU, - LWTRANCHE_SERIAL_SLRU, - LWTRANCHE_SUBTRANS_SLRU, - LWTRANCHE_XACT_SLRU, - LWTRANCHE_PARALLEL_VACUUM_DSA, - LWTRANCHE_AIO_URING_COMPLETION, + /* + * LWTRANCHE_INVALID is an unused value that only exists to initialize the + * rest of the tranches to appropriate values. + */ + LWTRANCHE_INVALID = NUM_INDIVIDUAL_LWLOCKS - 1, + +#define PG_LWLOCK(id, name) +#define PG_LWLOCKTRANCHE(id, name) LWTRANCHE_##id, +#include "storage/lwlocklist.h" +#undef PG_LWLOCK +#undef PG_LWLOCKTRANCHE + LWTRANCHE_FIRST_USER_DEFINED, } BuiltinTrancheIds; diff --git a/src/include/storage/lwlocklist.h b/src/include/storage/lwlocklist.h index a9681738146e1..208d2e3a8ed9e 100644 --- a/src/include/storage/lwlocklist.h +++ b/src/include/storage/lwlocklist.h @@ -2,9 +2,10 @@ * * lwlocklist.h * - * The predefined LWLock list is kept in its own source file for use by - * automatic tools. The exact representation of a keyword is determined by - * the PG_LWLOCK macro, which is not defined in this file; it can be + * The list of predefined LWLocks and built-in LWLock tranches is kept in + * its own source file for use by automatic tools. The exact + * representation of a keyword is determined by the PG_LWLOCK and + * PG_LWLOCKTRANCHE macros, which are not defined in this file; they can be * defined by the caller for special purposes. * * Also, generate-lwlocknames.pl processes this file to create lwlocknames.h. @@ -84,3 +85,53 @@ PG_LWLOCK(50, DSMRegistry) PG_LWLOCK(51, InjectionPoint) PG_LWLOCK(52, SerialControl) PG_LWLOCK(53, AioWorkerSubmissionQueue) + +/* + * There also exist several built-in LWLock tranches. As with the predefined + * LWLocks, be sure to update the WaitEventLWLock section of + * src/backend/utils/activity/wait_event_names.txt when modifying this list. + * + * Note that the IDs here (the first value) don't include the LWTRANCHE_ + * prefix. It's added elsewhere. + */ +PG_LWLOCKTRANCHE(XACT_BUFFER, XactBuffer) +PG_LWLOCKTRANCHE(COMMITTS_BUFFER, CommitTsBuffer) +PG_LWLOCKTRANCHE(SUBTRANS_BUFFER, SubtransBuffer) +PG_LWLOCKTRANCHE(MULTIXACTOFFSET_BUFFER, MultiXactOffsetBuffer) +PG_LWLOCKTRANCHE(MULTIXACTMEMBER_BUFFER, MultiXactMemberBuffer) +PG_LWLOCKTRANCHE(NOTIFY_BUFFER, NotifyBuffer) +PG_LWLOCKTRANCHE(SERIAL_BUFFER, SerialBuffer) +PG_LWLOCKTRANCHE(WAL_INSERT, WALInsert) +PG_LWLOCKTRANCHE(BUFFER_CONTENT, BufferContent) +PG_LWLOCKTRANCHE(REPLICATION_ORIGIN_STATE, ReplicationOriginState) +PG_LWLOCKTRANCHE(REPLICATION_SLOT_IO, ReplicationSlotIO) +PG_LWLOCKTRANCHE(LOCK_FASTPATH, LockFastPath) +PG_LWLOCKTRANCHE(BUFFER_MAPPING, BufferMapping) +PG_LWLOCKTRANCHE(LOCK_MANAGER, LockManager) +PG_LWLOCKTRANCHE(PREDICATE_LOCK_MANAGER, PredicateLockManager) +PG_LWLOCKTRANCHE(PARALLEL_HASH_JOIN, ParallelHashJoin) +PG_LWLOCKTRANCHE(PARALLEL_BTREE_SCAN, ParallelBtreeScan) +PG_LWLOCKTRANCHE(PARALLEL_QUERY_DSA, ParallelQueryDSA) +PG_LWLOCKTRANCHE(PER_SESSION_DSA, PerSessionDSA) +PG_LWLOCKTRANCHE(PER_SESSION_RECORD_TYPE, PerSessionRecordType) +PG_LWLOCKTRANCHE(PER_SESSION_RECORD_TYPMOD, PerSessionRecordTypmod) +PG_LWLOCKTRANCHE(SHARED_TUPLESTORE, SharedTupleStore) +PG_LWLOCKTRANCHE(SHARED_TIDBITMAP, SharedTidBitmap) +PG_LWLOCKTRANCHE(PARALLEL_APPEND, ParallelAppend) +PG_LWLOCKTRANCHE(PER_XACT_PREDICATE_LIST, PerXactPredicateList) +PG_LWLOCKTRANCHE(PGSTATS_DSA, PgStatsDSA) +PG_LWLOCKTRANCHE(PGSTATS_HASH, PgStatsHash) +PG_LWLOCKTRANCHE(PGSTATS_DATA, PgStatsData) +PG_LWLOCKTRANCHE(LAUNCHER_DSA, LogicalRepLauncherDSA) +PG_LWLOCKTRANCHE(LAUNCHER_HASH, LogicalRepLauncherHash) +PG_LWLOCKTRANCHE(DSM_REGISTRY_DSA, DSMRegistryDSA) +PG_LWLOCKTRANCHE(DSM_REGISTRY_HASH, DSMRegistryHash) +PG_LWLOCKTRANCHE(COMMITTS_SLRU, CommitTsSLRU) +PG_LWLOCKTRANCHE(MULTIXACTOFFSET_SLRU, MultiXactOffsetSLRU) +PG_LWLOCKTRANCHE(MULTIXACTMEMBER_SLRU, MultiXactMemberSLRU) +PG_LWLOCKTRANCHE(NOTIFY_SLRU, NotifySLRU) +PG_LWLOCKTRANCHE(SERIAL_SLRU, SerialSLRU) +PG_LWLOCKTRANCHE(SUBTRANS_SLRU, SubtransSLRU) +PG_LWLOCKTRANCHE(XACT_SLRU, XactSLRU) +PG_LWLOCKTRANCHE(PARALLEL_VACUUM_DSA, ParallelVacuumDSA) +PG_LWLOCKTRANCHE(AIO_URING_COMPLETION, AioUringCompletion) diff --git a/src/include/utils/catcache.h b/src/include/utils/catcache.h index 277ec33c00bac..00808e23f49b8 100644 --- a/src/include/utils/catcache.h +++ b/src/include/utils/catcache.h @@ -87,6 +87,14 @@ typedef struct catcache typedef struct catctup { + /* + * Each tuple in a cache is a member of a dlist that stores the elements + * of its hash bucket. We keep each dlist in LRU order to speed repeated + * lookups. Keep the dlist_node field first so that Valgrind understands + * the struct is reachable. + */ + dlist_node cache_elem; /* list member of per-bucket list */ + int ct_magic; /* for identifying CatCTup entries */ #define CT_MAGIC 0x57261502 @@ -98,13 +106,6 @@ typedef struct catctup */ Datum keys[CATCACHE_MAXKEYS]; - /* - * Each tuple in a cache is a member of a dlist that stores the elements - * of its hash bucket. We keep each dlist in LRU order to speed repeated - * lookups. - */ - dlist_node cache_elem; /* list member of per-bucket list */ - /* * A tuple marked "dead" must not be returned by subsequent searches. * However, it won't be physically deleted from the cache until its @@ -158,13 +159,17 @@ typedef struct catctup */ typedef struct catclist { + /* + * Keep the dlist_node field first so that Valgrind understands the struct + * is reachable. + */ + dlist_node cache_elem; /* list member of per-catcache list */ + int cl_magic; /* for identifying CatCList entries */ #define CL_MAGIC 0x52765103 uint32 hash_value; /* hash value for lookup keys */ - dlist_node cache_elem; /* list member of per-catcache list */ - /* * Lookup keys for the entry, with the first nkeys elements being valid. * All by-reference are separately allocated. diff --git a/src/include/utils/memdebug.h b/src/include/utils/memdebug.h index 7309271834b9f..80692dcef9382 100644 --- a/src/include/utils/memdebug.h +++ b/src/include/utils/memdebug.h @@ -29,6 +29,7 @@ #define VALGRIND_MEMPOOL_ALLOC(context, addr, size) do {} while (0) #define VALGRIND_MEMPOOL_FREE(context, addr) do {} while (0) #define VALGRIND_MEMPOOL_CHANGE(context, optr, nptr, size) do {} while (0) +#define VALGRIND_MEMPOOL_TRIM(context, addr, size) do {} while (0) #endif diff --git a/src/include/utils/palloc.h b/src/include/utils/palloc.h index e1b42267b22aa..039b9cba61a32 100644 --- a/src/include/utils/palloc.h +++ b/src/include/utils/palloc.h @@ -133,6 +133,8 @@ MemoryContextSwitchTo(MemoryContext context) /* Registration of memory context reset/delete callbacks */ extern void MemoryContextRegisterResetCallback(MemoryContext context, MemoryContextCallback *cb); +extern void MemoryContextUnregisterResetCallback(MemoryContext context, + MemoryContextCallback *cb); /* * These are like standard strdup() except the copied string is diff --git a/src/include/utils/pgstat_internal.h b/src/include/utils/pgstat_internal.h index d5557e6e998cd..6cf00008f6333 100644 --- a/src/include/utils/pgstat_internal.h +++ b/src/include/utils/pgstat_internal.h @@ -295,18 +295,11 @@ typedef struct PgStat_KindInfo * * Returns true if some of the stats could not be flushed, due to lock * contention for example. Optional. - */ - bool (*flush_static_cb) (bool nowait); - - /* - * For fixed-numbered or variable-numbered statistics: Check for pending - * stats in need of flush with flush_static_cb, when these do not use - * PgStat_EntryRef->pending. * - * Returns true if there are any stats pending for flush, triggering - * flush_static_cb. Optional. + * "pgstat_report_fixed" needs to be set to trigger the flush of pending + * stats. */ - bool (*have_static_pending_cb) (void); + bool (*flush_static_cb) (bool nowait); /* * For fixed-numbered statistics: Reset All. @@ -627,7 +620,6 @@ extern void pgstat_archiver_snapshot_cb(void); extern bool pgstat_flush_backend(bool nowait, bits32 flags); extern bool pgstat_backend_flush_cb(bool nowait); -extern bool pgstat_backend_have_pending_cb(void); extern void pgstat_backend_reset_timestamp_cb(PgStatShared_Common *header, TimestampTz ts); @@ -676,7 +668,6 @@ extern bool pgstat_function_flush_cb(PgStat_EntryRef *entry_ref, bool nowait); extern void pgstat_flush_io(bool nowait); -extern bool pgstat_io_have_pending_cb(void); extern bool pgstat_io_flush_cb(bool nowait); extern void pgstat_io_init_shmem_cb(void *stats); extern void pgstat_io_reset_all_cb(TimestampTz ts); @@ -738,7 +729,6 @@ extern PgStatShared_Common *pgstat_init_entry(PgStat_Kind kind, * Functions in pgstat_slru.c */ -extern bool pgstat_slru_have_pending_cb(void); extern bool pgstat_slru_flush_cb(bool nowait); extern void pgstat_slru_init_shmem_cb(void *stats); extern void pgstat_slru_reset_all_cb(TimestampTz ts); @@ -750,7 +740,6 @@ extern void pgstat_slru_snapshot_cb(void); */ extern void pgstat_wal_init_backend_cb(void); -extern bool pgstat_wal_have_pending_cb(void); extern bool pgstat_wal_flush_cb(bool nowait); extern void pgstat_wal_init_shmem_cb(void *stats); extern void pgstat_wal_reset_all_cb(TimestampTz ts); @@ -778,8 +767,23 @@ extern void pgstat_create_transactional(PgStat_Kind kind, Oid dboid, uint64 obji * Variables in pgstat.c */ -extern PGDLLIMPORT PgStat_LocalState pgStatLocal; +/* + * Track if *any* pending fixed-numbered statistics should be flushed to + * shared memory. + * + * This flag can be switched to true by fixed-numbered statistics to let + * pgstat_report_stat() know if it needs to go through one round of + * reports, calling flush_static_cb for each fixed-numbered statistics + * kind. When this flag is not set, pgstat_report_stat() is able to do + * a fast exit, knowing that there are no pending fixed-numbered statistics. + * + * Statistics callbacks should never reset this flag; pgstat_report_stat() + * is in charge of doing that. + */ +extern PGDLLIMPORT bool pgstat_report_fixed; +/* Backend-local stats state */ +extern PGDLLIMPORT PgStat_LocalState pgStatLocal; /* * Implementation of inline functions declared above. diff --git a/src/include/utils/pgstat_kind.h b/src/include/utils/pgstat_kind.h index f44169fd5a3c7..eb5f0b3ae6db7 100644 --- a/src/include/utils/pgstat_kind.h +++ b/src/include/utils/pgstat_kind.h @@ -18,7 +18,7 @@ /* Range of IDs allowed, for built-in and custom kinds */ #define PGSTAT_KIND_MIN 1 /* Minimum ID allowed */ -#define PGSTAT_KIND_MAX 256 /* Maximum ID allowed */ +#define PGSTAT_KIND_MAX 32 /* Maximum ID allowed */ /* use 0 for INVALID, to catch zero-initialized data */ #define PGSTAT_KIND_INVALID 0 @@ -46,7 +46,7 @@ /* Custom stats kinds */ /* Range of IDs allowed for custom stats kinds */ -#define PGSTAT_KIND_CUSTOM_MIN 128 +#define PGSTAT_KIND_CUSTOM_MIN 24 #define PGSTAT_KIND_CUSTOM_MAX PGSTAT_KIND_MAX #define PGSTAT_KIND_CUSTOM_SIZE (PGSTAT_KIND_CUSTOM_MAX - PGSTAT_KIND_CUSTOM_MIN + 1) @@ -55,7 +55,7 @@ * development and have not reserved their own unique kind ID yet. See: * https://wiki.postgresql.org/wiki/CustomCumulativeStats */ -#define PGSTAT_KIND_EXPERIMENTAL 128 +#define PGSTAT_KIND_EXPERIMENTAL 24 static inline bool pgstat_is_kind_builtin(PgStat_Kind kind) diff --git a/src/interfaces/ecpg/ecpglib/prepare.c b/src/interfaces/ecpg/ecpglib/prepare.c index dd6fd1fe7f407..06f0135813b37 100644 --- a/src/interfaces/ecpg/ecpglib/prepare.c +++ b/src/interfaces/ecpg/ecpglib/prepare.c @@ -603,7 +603,10 @@ ecpg_auto_prepare(int lineno, const char *connection_name, const int compat, cha prep = ecpg_find_prepared_statement(stmtID, con, NULL); /* This prepared name doesn't exist on this connection. */ if (!prep && !prepare_common(lineno, con, stmtID, query)) + { + ecpg_free(*name); return false; + } } else @@ -619,11 +622,17 @@ ecpg_auto_prepare(int lineno, const char *connection_name, const int compat, cha return false; if (!ECPGprepare(lineno, connection_name, 0, stmtID, query)) + { + ecpg_free(*name); return false; + } entNo = AddStmtToCache(lineno, stmtID, connection_name, compat, query); if (entNo < 0) + { + ecpg_free(*name); return false; + } } /* increase usage counter */ diff --git a/src/interfaces/libpq-oauth/Makefile b/src/interfaces/libpq-oauth/Makefile index 270fc0cf2d9d9..682f17413b3a4 100644 --- a/src/interfaces/libpq-oauth/Makefile +++ b/src/interfaces/libpq-oauth/Makefile @@ -24,7 +24,7 @@ NAME = pq-oauth-$(MAJORVERSION) override shlib := lib$(NAME)$(DLSUFFIX) override stlib := libpq-oauth.a -override CPPFLAGS := -I$(libpq_srcdir) -I$(top_builddir)/src/port $(LIBCURL_CPPFLAGS) $(CPPFLAGS) +override CPPFLAGS := -I$(libpq_srcdir) -I$(top_builddir)/src/port $(CPPFLAGS) $(LIBCURL_CPPFLAGS) OBJS = \ $(WIN32RES) diff --git a/src/interfaces/libpq/Makefile b/src/interfaces/libpq/Makefile index 47d6781150944..da6650066d46e 100644 --- a/src/interfaces/libpq/Makefile +++ b/src/interfaces/libpq/Makefile @@ -24,7 +24,7 @@ NAME= pq SO_MAJOR_VERSION= 5 SO_MINOR_VERSION= $(MAJORVERSION) -override CPPFLAGS := -I$(srcdir) $(CPPFLAGS) -I$(top_builddir)/src/port -I$(top_srcdir)/src/port +override CPPFLAGS := -I$(srcdir) -I$(top_builddir)/src/port -I$(top_srcdir)/src/port $(CPPFLAGS) ifneq ($(PORTNAME), win32) override CFLAGS += $(PTHREAD_CFLAGS) endif diff --git a/src/interfaces/libpq/fe-cancel.c b/src/interfaces/libpq/fe-cancel.c index 65517c5703bca..c872a0267f089 100644 --- a/src/interfaces/libpq/fe-cancel.c +++ b/src/interfaces/libpq/fe-cancel.c @@ -379,7 +379,24 @@ PQgetCancel(PGconn *conn) /* Check that we have received a cancellation key */ if (conn->be_cancel_key_len == 0) - return NULL; + { + /* + * In case there is no cancel key, return an all-zero PGcancel object. + * Actually calling PQcancel on this will fail, but we allow creating + * the PGcancel object anyway. Arguably it would be better return NULL + * to indicate that cancellation is not possible, but there'd be no + * way for the caller to distinguish "out of memory" from "server did + * not send a cancel key". Also, this is how PGgetCancel() has always + * behaved, and if we changed it, some clients would stop working + * altogether with servers that don't support cancellation. (The + * modern PQcancelCreate() function returns a failed connection object + * instead.) + * + * The returned dummy object has cancel_pkt_len == 0; we check for + * that in PQcancel() to identify it as a dummy. + */ + return calloc(1, sizeof(PGcancel)); + } cancel_req_len = offsetof(CancelRequestPacket, cancelAuthCode) + conn->be_cancel_key_len; cancel = malloc(offsetof(PGcancel, cancel_req) + cancel_req_len); @@ -544,6 +561,15 @@ PQcancel(PGcancel *cancel, char *errbuf, int errbufsize) return false; } + if (cancel->cancel_pkt_len == 0) + { + /* This is a dummy PGcancel object, see PQgetCancel */ + strlcpy(errbuf, "PQcancel() -- no cancellation key received", errbufsize); + /* strlcpy probably doesn't change errno, but be paranoid */ + SOCK_ERRNO_SET(save_errno); + return false; + } + /* * We need to open a temporary connection to the postmaster. Do this with * only kernel calls. diff --git a/src/pl/plpgsql/src/pl_comp.c b/src/pl/plpgsql/src/pl_comp.c index ee961425a5b7e..f6976689a6927 100644 --- a/src/pl/plpgsql/src/pl_comp.c +++ b/src/pl/plpgsql/src/pl_comp.c @@ -177,6 +177,7 @@ plpgsql_compile_callback(FunctionCallInfo fcinfo, yyscan_t scanner; Datum prosrcdatum; char *proc_source; + char *proc_signature; HeapTuple typeTup; Form_pg_type typeStruct; PLpgSQL_variable *var; @@ -223,6 +224,9 @@ plpgsql_compile_callback(FunctionCallInfo fcinfo, plpgsql_check_syntax = forValidator; plpgsql_curr_compile = function; + /* format_procedure leaks memory, so run it in temp context */ + proc_signature = format_procedure(fcinfo->flinfo->fn_oid); + /* * All the permanent output of compilation (e.g. parse tree) is kept in a * per-function memory context, so it can be reclaimed easily. @@ -237,7 +241,7 @@ plpgsql_compile_callback(FunctionCallInfo fcinfo, ALLOCSET_DEFAULT_SIZES); plpgsql_compile_tmp_cxt = MemoryContextSwitchTo(func_cxt); - function->fn_signature = format_procedure(fcinfo->flinfo->fn_oid); + function->fn_signature = pstrdup(proc_signature); MemoryContextSetIdentifier(func_cxt, function->fn_signature); function->fn_oid = fcinfo->flinfo->fn_oid; function->fn_input_collation = fcinfo->fncollation; @@ -1673,6 +1677,11 @@ plpgsql_parse_wordrowtype(char *ident) { Oid classOid; Oid typOid; + TypeName *typName; + MemoryContext oldCxt; + + /* Avoid memory leaks in long-term function context */ + oldCxt = MemoryContextSwitchTo(plpgsql_compile_tmp_cxt); /* * Look up the relation. Note that because relation rowtypes have the @@ -1695,9 +1704,12 @@ plpgsql_parse_wordrowtype(char *ident) errmsg("relation \"%s\" does not have a composite type", ident))); + typName = makeTypeName(ident); + + MemoryContextSwitchTo(oldCxt); + /* Build and return the row type struct */ - return plpgsql_build_datatype(typOid, -1, InvalidOid, - makeTypeName(ident)); + return plpgsql_build_datatype(typOid, -1, InvalidOid, typName); } /* ---------- @@ -1711,6 +1723,7 @@ plpgsql_parse_cwordrowtype(List *idents) Oid classOid; Oid typOid; RangeVar *relvar; + TypeName *typName; MemoryContext oldCxt; /* @@ -1733,11 +1746,12 @@ plpgsql_parse_cwordrowtype(List *idents) errmsg("relation \"%s\" does not have a composite type", relvar->relname))); + typName = makeTypeNameFromNameList(idents); + MemoryContextSwitchTo(oldCxt); /* Build and return the row type struct */ - return plpgsql_build_datatype(typOid, -1, InvalidOid, - makeTypeNameFromNameList(idents)); + return plpgsql_build_datatype(typOid, -1, InvalidOid, typName); } /* @@ -1952,6 +1966,8 @@ plpgsql_build_recfield(PLpgSQL_rec *rec, const char *fldname) * origtypname is the parsed form of what the user wrote as the type name. * It can be NULL if the type could not be a composite type, or if it was * identified by OID to begin with (e.g., it's a function argument type). + * origtypname is in short-lived storage and must be copied if we choose + * to incorporate it into the function's parse tree. */ PLpgSQL_type * plpgsql_build_datatype(Oid typeOid, int32 typmod, @@ -2070,7 +2086,7 @@ build_datatype(HeapTuple typeTup, int32 typmod, errmsg("type %s is not composite", format_type_be(typ->typoid)))); - typ->origtypname = origtypname; + typ->origtypname = copyObject(origtypname); typ->tcache = typentry; typ->tupdesc_id = typentry->tupDesc_identifier; } diff --git a/src/pl/plpgsql/src/pl_gram.y b/src/pl/plpgsql/src/pl_gram.y index 7b672ea5179a6..17568d82554d2 100644 --- a/src/pl/plpgsql/src/pl_gram.y +++ b/src/pl/plpgsql/src/pl_gram.y @@ -3853,6 +3853,7 @@ parse_datatype(const char *string, int location, yyscan_t yyscanner) int32 typmod; sql_error_callback_arg cbarg; ErrorContextCallback syntax_errcontext; + MemoryContext oldCxt; cbarg.location = location; cbarg.yyscanner = yyscanner; @@ -3862,9 +3863,14 @@ parse_datatype(const char *string, int location, yyscan_t yyscanner) syntax_errcontext.previous = error_context_stack; error_context_stack = &syntax_errcontext; - /* Let the main parser try to parse it under standard SQL rules */ + /* + * Let the main parser try to parse it under standard SQL rules. The + * parser leaks memory, so run it in temp context. + */ + oldCxt = MemoryContextSwitchTo(plpgsql_compile_tmp_cxt); typeName = typeStringToTypeName(string, NULL); typenameTypeIdAndMod(NULL, typeName, &type_id, &typmod); + MemoryContextSwitchTo(oldCxt); /* Restore former ereport callback */ error_context_stack = syntax_errcontext.previous; diff --git a/src/pl/plpython/Makefile b/src/pl/plpython/Makefile index f959083a0bdec..25f295c3709e2 100644 --- a/src/pl/plpython/Makefile +++ b/src/pl/plpython/Makefile @@ -11,7 +11,7 @@ ifeq ($(PORTNAME), win32) override python_libspec = endif -override CPPFLAGS := -I. -I$(srcdir) $(python_includespec) $(CPPFLAGS) +override CPPFLAGS := -I. -I$(srcdir) $(CPPFLAGS) $(python_includespec) rpathdir = $(python_libdir) diff --git a/src/pl/tcl/Makefile b/src/pl/tcl/Makefile index ea52a2efc229d..dd57f7d694c82 100644 --- a/src/pl/tcl/Makefile +++ b/src/pl/tcl/Makefile @@ -11,7 +11,7 @@ top_builddir = ../../.. include $(top_builddir)/src/Makefile.global -override CPPFLAGS := -I. -I$(srcdir) $(TCL_INCLUDE_SPEC) $(CPPFLAGS) +override CPPFLAGS := -I. -I$(srcdir) $(CPPFLAGS) $(TCL_INCLUDE_SPEC) # On Windows, we don't link directly with the Tcl library; see below ifneq ($(PORTNAME), win32) diff --git a/src/test/modules/injection_points/injection_stats.c b/src/test/modules/injection_points/injection_stats.c index 14903c629e0d1..e3947b23ba573 100644 --- a/src/test/modules/injection_points/injection_stats.c +++ b/src/test/modules/injection_points/injection_stats.c @@ -59,7 +59,7 @@ static const PgStat_KindInfo injection_stats = { /* * Kind ID reserved for statistics of injection points. */ -#define PGSTAT_KIND_INJECTION 129 +#define PGSTAT_KIND_INJECTION 25 /* Track if stats are loaded */ static bool inj_stats_loaded = false; diff --git a/src/test/modules/injection_points/injection_stats_fixed.c b/src/test/modules/injection_points/injection_stats_fixed.c index 3d0c01bdd05ab..bc54c79d190b9 100644 --- a/src/test/modules/injection_points/injection_stats_fixed.c +++ b/src/test/modules/injection_points/injection_stats_fixed.c @@ -64,7 +64,7 @@ static const PgStat_KindInfo injection_stats_fixed = { /* * Kind ID reserved for statistics of injection points. */ -#define PGSTAT_KIND_INJECTION_FIXED 130 +#define PGSTAT_KIND_INJECTION_FIXED 26 /* Track if fixed-numbered stats are loaded */ static bool inj_fixed_loaded = false; diff --git a/src/test/perl/PostgreSQL/Test/Cluster.pm b/src/test/perl/PostgreSQL/Test/Cluster.pm index 61f68e0cc2e51..35413f140198b 100644 --- a/src/test/perl/PostgreSQL/Test/Cluster.pm +++ b/src/test/perl/PostgreSQL/Test/Cluster.pm @@ -304,6 +304,7 @@ sub is_alive my $ret = PostgreSQL::Test::Utils::system_log( 'pg_isready', + '--timeout' => $PostgreSQL::Test::Utils::timeout_default, '--host' => $self->host, '--port' => $self->port); diff --git a/src/test/recovery/t/013_crash_restart.pl b/src/test/recovery/t/013_crash_restart.pl index debfa635c36fe..4c5af018ee44e 100644 --- a/src/test/recovery/t/013_crash_restart.pl +++ b/src/test/recovery/t/013_crash_restart.pl @@ -228,6 +228,13 @@ 'before-orderly-restart', 'can still write after crash restart'); +# Confirm that the logical replication launcher, a background worker +# without the never-restart flag, has also restarted successfully. +is($node->poll_query_until('postgres', + "SELECT count(*) = 1 FROM pg_stat_activity WHERE backend_type = 'logical replication launcher'"), + '1', + 'logical replication launcher restarted after crash'); + # Just to be sure, check that an orderly restart now still works $node->restart(); diff --git a/src/test/regress/expected/foreign_key.out b/src/test/regress/expected/foreign_key.out index f9bd252444f53..dc541d61adfa5 100644 --- a/src/test/regress/expected/foreign_key.out +++ b/src/test/regress/expected/foreign_key.out @@ -1750,7 +1750,7 @@ Indexes: Referenced by: TABLE "fk_partitioned_fk" CONSTRAINT "fk_partitioned_fk_a_b_fkey" FOREIGN KEY (a, b) REFERENCES fk_notpartitioned_pk(a, b) --- Check the exsting FK trigger +-- Check the existing FK trigger SELECT conname, tgrelid::regclass as tgrel, regexp_replace(tgname, '[0-9]+', 'N') as tgname, tgtype FROM pg_trigger t JOIN pg_constraint c ON (t.tgconstraint = c.oid) WHERE tgrelid IN (SELECT relid FROM pg_partition_tree('fk_partitioned_fk'::regclass) diff --git a/src/test/regress/expected/publication.out b/src/test/regress/expected/publication.out index 3a2eacd793f70..1ec3fa34a2d5a 100644 --- a/src/test/regress/expected/publication.out +++ b/src/test/regress/expected/publication.out @@ -1934,3 +1934,24 @@ RESET client_min_messages; RESET SESSION AUTHORIZATION; DROP ROLE regress_publication_user, regress_publication_user2; DROP ROLE regress_publication_user_dummy; +-- stage objects for pg_dump tests +CREATE SCHEMA pubme CREATE TABLE t0 (c int, d int) CREATE TABLE t1 (c int); +CREATE SCHEMA pubme2 CREATE TABLE t0 (c int, d int); +SET client_min_messages = 'ERROR'; +CREATE PUBLICATION dump_pub_qual_1ct FOR + TABLE ONLY pubme.t0 (c, d) WHERE (c > 0); +CREATE PUBLICATION dump_pub_qual_2ct FOR + TABLE ONLY pubme.t0 (c) WHERE (c > 0), + TABLE ONLY pubme.t1 (c); +CREATE PUBLICATION dump_pub_nsp_1ct FOR + TABLES IN SCHEMA pubme; +CREATE PUBLICATION dump_pub_nsp_2ct FOR + TABLES IN SCHEMA pubme, + TABLES IN SCHEMA pubme2; +CREATE PUBLICATION dump_pub_all FOR + TABLE ONLY pubme.t0, + TABLE ONLY pubme.t1 WHERE (c < 0), + TABLES IN SCHEMA pubme, + TABLES IN SCHEMA pubme2 + WITH (publish_via_partition_root = true); +RESET client_min_messages; diff --git a/src/test/regress/sql/foreign_key.sql b/src/test/regress/sql/foreign_key.sql index cfcecb4e911ad..39174ad1eb9a0 100644 --- a/src/test/regress/sql/foreign_key.sql +++ b/src/test/regress/sql/foreign_key.sql @@ -1296,7 +1296,7 @@ UPDATE fk_notpartitioned_pk SET b = 2504 WHERE a = 2500; -- check psql behavior \d fk_notpartitioned_pk --- Check the exsting FK trigger +-- Check the existing FK trigger SELECT conname, tgrelid::regclass as tgrel, regexp_replace(tgname, '[0-9]+', 'N') as tgname, tgtype FROM pg_trigger t JOIN pg_constraint c ON (t.tgconstraint = c.oid) WHERE tgrelid IN (SELECT relid FROM pg_partition_tree('fk_partitioned_fk'::regclass) diff --git a/src/test/regress/sql/publication.sql b/src/test/regress/sql/publication.sql index c9e309190dfa6..2585f08318150 100644 --- a/src/test/regress/sql/publication.sql +++ b/src/test/regress/sql/publication.sql @@ -1229,3 +1229,25 @@ RESET client_min_messages; RESET SESSION AUTHORIZATION; DROP ROLE regress_publication_user, regress_publication_user2; DROP ROLE regress_publication_user_dummy; + +-- stage objects for pg_dump tests +CREATE SCHEMA pubme CREATE TABLE t0 (c int, d int) CREATE TABLE t1 (c int); +CREATE SCHEMA pubme2 CREATE TABLE t0 (c int, d int); +SET client_min_messages = 'ERROR'; +CREATE PUBLICATION dump_pub_qual_1ct FOR + TABLE ONLY pubme.t0 (c, d) WHERE (c > 0); +CREATE PUBLICATION dump_pub_qual_2ct FOR + TABLE ONLY pubme.t0 (c) WHERE (c > 0), + TABLE ONLY pubme.t1 (c); +CREATE PUBLICATION dump_pub_nsp_1ct FOR + TABLES IN SCHEMA pubme; +CREATE PUBLICATION dump_pub_nsp_2ct FOR + TABLES IN SCHEMA pubme, + TABLES IN SCHEMA pubme2; +CREATE PUBLICATION dump_pub_all FOR + TABLE ONLY pubme.t0, + TABLE ONLY pubme.t1 WHERE (c < 0), + TABLES IN SCHEMA pubme, + TABLES IN SCHEMA pubme2 + WITH (publish_via_partition_root = true); +RESET client_min_messages; diff --git a/src/test/subscription/t/035_conflicts.pl b/src/test/subscription/t/035_conflicts.pl index 7458d7fba7e9a..976d53a870e5e 100644 --- a/src/test/subscription/t/035_conflicts.pl +++ b/src/test/subscription/t/035_conflicts.pl @@ -228,6 +228,11 @@ # Disable the subscription $node_A->psql('postgres', "ALTER SUBSCRIPTION $subname_AB DISABLE;"); +# Wait for the apply worker to stop +$node_A->poll_query_until('postgres', + "SELECT count(*) = 0 FROM pg_stat_activity WHERE backend_type = 'logical replication apply worker'" +); + # Enable retain_dead_tuples for disabled subscription ($cmdret, $stdout, $stderr) = $node_A->psql('postgres', "ALTER SUBSCRIPTION $subname_AB SET (retain_dead_tuples = true);"); @@ -278,6 +283,11 @@ # Disable the logical replication from node B to node A $node_A->safe_psql('postgres', "ALTER SUBSCRIPTION $subname_AB DISABLE"); +# Wait for the apply worker to stop +$node_A->poll_query_until('postgres', + "SELECT count(*) = 0 FROM pg_stat_activity WHERE backend_type = 'logical replication apply worker'" +); + $node_B->safe_psql('postgres', "UPDATE tab SET b = 3 WHERE a = 1;"); $node_A->safe_psql('postgres', "DELETE FROM tab WHERE a = 1;"); diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list index a8656419cb608..e6f2e93b2d6fa 100644 --- a/src/tools/pgindent/typedefs.list +++ b/src/tools/pgindent/typedefs.list @@ -2275,6 +2275,7 @@ PlanInvalItem PlanRowMark PlanState PlannedStmt +PlannedStmtOrigin PlannerGlobal PlannerInfo PlannerParamItem @@ -3756,6 +3757,7 @@ leafSegmentInfo leaf_item libpq_gettext_func libpq_source +libpqsrv_PGresult line_t lineno_t list_sort_comparator diff --git a/src/tools/valgrind.supp b/src/tools/valgrind.supp index 2ad5b81526d3f..3880007dfb3bb 100644 --- a/src/tools/valgrind.supp +++ b/src/tools/valgrind.supp @@ -194,3 +194,36 @@ Memcheck:Addr8 fun:pg_numa_touch_mem_if_required } + + +# Memory-leak suppressions +# Note that a suppression rule will silence complaints about memory blocks +# allocated in matching places, but it won't prevent "indirectly lost" +# complaints about blocks that are only reachable via the suppressed blocks. + +# Suppress complaints about stuff leaked during function cache loading. +# Both the PL/pgSQL and SQL-function parsing processes generate some cruft +# within the function's cache context, which doesn't seem worth the trouble +# to get rid of. Moreover, there are cases where CachedFunction structs +# are intentionally leaked because we're unsure if any fn_extra pointers +# remain. +{ + hide_function_cache_leaks + Memcheck:Leak + match-leak-kinds: definite,possible,indirect + + ... + fun:cached_function_compile +} + +# Suppress complaints about stuff leaked during TS dictionary loading. +# Not very much is typically lost there, and preventing it would +# require a risky API change for TS tmplinit functions. +{ + hide_ts_dictionary_leaks + Memcheck:Leak + match-leak-kinds: definite,possible,indirect + + ... + fun:lookup_ts_dictionary_cache +} pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy